The Complexity of Ageism : A Proposed Typology

Comparing data from a contemporary (2002) study and a study from 1984, it was found that, now as then, people display seemingly contradictory constellations of attitudes toward old people. Large proportions of respondents, now as then, advocate more influence and space for the 65+ group, at the same time as many feel that no one in the parliament should be above the age of 65. This contradiction becomes intelligible when the conceptions of and behavioral dispositions toward old people are combined in a new proposed typology of ageism, which is the result of a study conducted in 2002 and reported in this article. This new typology includes the Pitying Positive, the No Fuzz, the Consistently Negative and the Consistently Positive. These types are empirically described, and use of the typology is exemplified by focusing on ageist attitudes toward parliament membership.

Background and Aim of the Study When Butler (1969) introduced the term ageism, which he described as prejudices against other age groups, social gerontology gained a new and stimulating concept, even though other kindred concepts had been used earlier to target similar phenomena. In our early Swedish research, the theoretical point of departure was the traditional concept of attitudes. Based on pioneer writings on this subject (e.g., Thurstone 1929; Kretch & Crutchfield 1948;Green 1954;Katz 1960), we decided to include the three classic dimensions in our definition of attitudes toward old people and to focus on them: 1) the emotional or evaluative dimension, 2) the cognitive or "knowledge" dimension and 3) the behavioral disposition dimension (Tornstam 1984). Adding to the several definitions of ageism, our argument is that ageism may be fruitfully described in terms of these three dimensions -how people feel about and evaluate old people, conceptions of old people, and the disposition to behave or actual behavior in relation to old people. When Butler (2001), in his contemporary writings, describes what ageism is, he certainly touches upon the content of the three abovementioned dimensions, just as do Kite and Smith Wagner (2004), who explicitly state that ageist attitudes are best understood as a constellation of these three dimensions. Even Palmore's (1999) definition of ageism, with the distinction between prejudices and discrimination, can accommodate both prejudiced conceptions as well as dispositions to behave in a discriminating way. The focus in the present study is however more on the social psychological level and on behavior dispositions than on factual discriminatory acts.
When we, back in the 1980s, carried out studies on attitudes toward old people, with particular focus on behavioral dispositions, we found inconsistencies in the attitudes of Swedish respondents. At the same time as 97 percent of respondents thought that one important duty of old people is to share their experience with younger people, no less than 61 percent felt that no politician in parliament should be older than 65 (Tornstam 1984).
Stimulated by these and other related findings, the SOM Institute (Society Opinion Mass Media) at Göteborg University used some of our original survey items in one of their surveys in 2002. Despite the time difference of almost 20 years, the SOM study reveals a similar pattern in which 87 percent of Swedish respondents would like to see more places where young and old can meet, while 26 percent of the same group think it is a good or very good idea that no politician in parliament should be older than 65 years of age. Without diving into any discussion on the deeper meaning of the two items, it can be observed that a generally positive attitude toward the participation and visibility of old people in society seems to be combined with a quite negative attitude toward their participation and visibility in parliament, which we will refer to below as "parliament ageism".
The fact that the proportion of respondents negative to parliament membership for persons 65+ seems to have decreased from 61 percent to 26 percent during the 20-year period is at least partly due to methodological differences. In the 1984 study, respondents were only given an "agree" or "not agree" alternative, while respondents in the 2002 study were given a scale with five response alternatives ranging from "very good suggestion" to "very bad suggestion". In the 1984 study, 61 percent agreed with the statement that no one in the parliament should be older than 65, while in the 2002 study, 26 percent regarded this a "good" or "very good" suggestion. If the middle "neither good nor bad suggestion" is added, this figure increases to 72 percent.
It may be worth mentioning that the "parliament ageism" described above also exists among established politicians. A Swedish member of the European Parliament, Nils Lundgren (68 years of age), was recently criticized by some Swedish opponent politicians, who recommended that he withdraw to a lifestyle more appropriate for pensioners. Lundgren (2005) defended himself in a newspaper article by pointing out that some of the most influential politicians in history have been well above the age of 65. For instance, Winston Churchill returned as Prime Minister at 77 and withdrew when he was 81, and Ronald Reagan was 74 when elected to his second term as President of the United States. It is just as interesting that the age of 68 may still be considered an argument for withdrawal from parliament as it is that a parliament member feels the need to defend his position in the above manner.
When talking about ageism, the attitude toward 65+ membership in parliament is crucial. The perceived non-legitimacy of parliament membership may be seen a major indicator of ageism, and as a manifest behavioral inclination to support a corresponding legal regulation. Such a regulation, however, may by many be difficult to understand as ageism. Due to accustomedness with a socially constructed norm system, some of us feel it is right that old people leave room for the young. Many of the rules governing today's labor market imply this, even if the European Union is now trying to encourage member countries to regard compulsory retirement at a certain age as illegal discrimination -ageism. Yet there is some reluctance with regard to this new rule, as many of us still regard it as "natural" to stop working at 65, and consequently natural that old members of parliament should leave room for younger persons, just as did the above-mentioned political opponents to the 68-year-old Swedish member of the European Parliament.
Consider, however, what would happen if 26 percent of the population were to advocate a rule that no member of the parliament should be a woman: we would react with a riot (as did many women and some men in the past). If we free ourselves from our accustomedness with the hidden ageism inherent in some norms of exclusion, it may seem defendable, for the moment, to label those who have the above-mentioned attitude as "parliament ageists".
In any case, this persistency of contradictory attitudes has prompted us to take a closer look at this pattern. From a theoretical point of view, our approach is to consider ageism as a complex phenomenon rather than as a simple, one-dimensional phenomenon. Part of what characterizes ageism (or attitudes toward old people) is probably that it is comprised of seemingly contradictory discourses. As suggested in the present study, a focus on these discourses could produce a fuller understanding of the anatomy of ageism. And we may even find a hidden logic behind some apparent contradictions.
Thus, the major aim of this study has been to further the conceptual discussion on ageism by suggesting a new ageism typology based on combinations of the cognitive and behavioral disposition components. This constellation has been forced to the fore by the above-mentioned contradictions. At the same time, this means that, in this study, we have left the emotional component of attitudes aside for future consideration. Unfortunately, the SOM study, on which we build, did not include any measurement of the emotional dimension.
The aim has also been to describe the profiles of respondents belonging to the different types in the typology and to exemplify use of the typology by applying it to the "parliamentary ageism" described above.

Material and Methods
The analysis below is based on the above-mentioned SOM Institute study conducted in 2002, which was a postal survey sent to a random sample of 3000 Swedes between 15 and 85 years of age. The net response rate after attrition due to death, emigration, etc., was 69 percent (1899 individuals). In the analyses below, the respondent base for the calculations will vary depending on internal dropouts and index calculation rules.

The Conceptions of Old People Scale
In order to tap the degree to which respondents have correct conceptions of old people or have negatively biased beliefs, the statements below were used. These statements were, in the SOM study, the remaining ones from our original study from the 1980s, which in part originated from Palmores (1977) first quiz on aging.
Respondents were asked to indicate whether each statement was correct or incorrect: • More than one third of all retirement pensioners live in their own private houses 1 (Statement correct. More than 50 percent of all retirement pensioners do so, 2 but 61 percent of the respondents marked the statement as incorrect). • Half of all retirement pensioners have impaired hearing 3 (Statement incorrect. Less than 30 percent have impaired hearing, 4 but 56 percent of the respondents marked the statement as correct).
• Almost half of all retirement pensioners feel bored and dissatisfied with their situation 5 (Statement incorrect. Around 85 percent are rather or very satisfied with their situation, 6 but 42 percent of the respondents marked the statement as correct).
• Almost three quarters of all retirement pensioners often see relatives and friends. 7 (Statement correct. Almost 75 percent of retirement pensioners see relatives and/or friends on at least a weekly basis, 8 but 44 percent of the respondents marked the statement as incorrect).
On the basis of the responses to the above statements, a simple additive conceptions or "knowledge scale" was constructed. Those who gave correct responses to all items scored the maximum of 4 on this scale (8 percent), while those who gave incorrect response to all items scored 0 and, at the same time, demonstrated a negative bias in their conceptions of old people (16 percent). Those who scored 4 have either really known or, perhaps, displayed a positive bias.

The General Pro-Old Behavioral Disposition Scale
In order to tap a general pro-old behavioral dispositional attitude, an index has been constructed based on responses to the following statements: • Increase the proportion of retirement pensioners among Swedish political decision makers! 9 Twenty-five percent of respondents considered this a good or very good suggestion. (Five response alternatives ranging from "very good suggestion" to "very bad suggestion").
• Elderly people are underrepresented in Swedish politics! 10 Twenty percent of respondents marked this statement as "absolutely right". (Ten response alternatives ranging from "absolutely wrong" to "absolutely right").
• The experiences of elderly people are not utilized in Sweden! 11 Twentyeight percent of respondents marked this statement as "absolutely right". (Ten response alternatives ranging from "absolutely wrong" to "absolutely right").
• Sweden would be a better country to live in if elderly people were given more authority! 12 Twelve percent of respondents marked this statement as "absolutely right". (Ten response alternatives ranging from "absolutely wrong" to "absolutely right"). An additive index was constructed based on the above items according to the rules given by Galtung (1969). 13 The Cronbach's Alpha for this scale is 0.70, which is generally considered satisfactory.

Ageist Attitude Toward 65+ Members of Parliament
In contrast to the above pro-old behavioral disposition scale, which targets a general, more unspecific behavioral disposition, we have also targeted the very specific and certainly important question of ageism in practice -respondents' opinions regarding eligibility for parliament. Should individuals 65+ be eligible for parliament or not?
The attitude toward politicians 65+ was measured using a single item: • No politician in parliament should be older than 65! Twenty-six percent of respondents regarded this as a good or a very good suggestion -an attitude we call "parliament ageism" (Five response alternatives ranging from "very good suggestion" to "very bad suggestion).
In addition to the above items, several single item measures, e.g. questions about work, education, etc., have been used in this study.

Intercorrelations between major measures
If the attitudes and the signs of ageism were one-dimensional and simple in their logic, we would expect to find strong and consistent correlations between conceptions of old people, general pro-old attitudes and opinions regarding 65+ membership in parliament. As seen in Table 1, the empirical reality does not support this assumption.
One of the correlations in Table 1 (top left cell) is quite contrary to what might be expected. The more correct (or positively biased) concep-13 The procedure includes a tricotomization of the items (with the series of response values turned in the same logical direction) before adding them up to an approximate Likert rank order scale. -.08 ** (n = 1 815) ** p < .01 tions the respondent has on the above-described conceptions scale, the lower the value on the general pro-old scale. In other words, negatively biased beliefs about old people are associated with high pro-old attitudes. The expectation was rather to find it the other way around. The obtained correlation is not strong, but could still indicate that the pro-old attitude is complex, since it to some extent seems based on negatively biased conceptions of old people.
Furthermore, the correlations that are in the expected direction are not particularly strong. Our data show that the higher value respondents have on the conceptions scale (accurate knowledge or positive bias), the less likely they are to recommend 65 as the upper age limit for members of parliament. This is, however, a very weak correlation of -.08.
In the same vein, even if our data show that a pro-old attitude makes it less likely to recommend 65 as the upper age limit for members of parliament, the correlation expressing this "self-evident" relationship is only -.22. A very much higher correlation would be expected for expressing a "self-evident" correlation between two variables measuring more or less the same thing. Together these observations indicate that the phenomenon of attitudes and ageism is more complex than sometimes believed.
14 The more knowledge about old people, the lower score on the pro-old behavioral disposition scale. 15 The higher value on the pro-old behavioral disposition scale and the better knowledge about old people, the less likely to recommend age 65 as upper limit in parliament. This indication is strengthened when we look at some correlates to the main measures. Table 2 shows, for example, correlations between educational level and all three major measures. As expected, we find a correlation indicating that those with higher education levels are better informed about old people and less likely to recommend a 65+ upper age limit for parliament members, but at the same time score lower on the general pro-old scale. As suggested above, this indicates that showing a positive attitude toward old people by scoring high on the general pro-old behavioral disposition scale may be something different from having positive conceptions of old people (fewer negatively biased beliefs) and being against a 65+ upper age limit for parliament members.
Pointing in the same direction, age differences show that the older the respondent, the more positive he/she scores on the general pro-old scale, but at the same time with a slight tendency to advocate a 65 age limit for parliament members. Our suggestion is that the general pro-old scale and the conceptions scale are measuring two dimensions that must be understood in non-additive ways. That is, these aspects of ageism cannot be added together to form a one-dimensional measure, but must be regarded as components of a constellation or interaction of attitudes.
This suggestion paves the way for experimenting with a typology based on these measures. This typology may offer new ways of understanding the seemingly contradictory response patterns.

A Typology of Attitudes
By means of dichotomizing the conceptions scale and the general pro-old scale, we have been able to construct a typology of "conceptions"/proold attitudes as shown in Table 3.
If the general pro-old scale and the conceptions scale consistently measured some common aspect of ageism/non-ageism, we would expect to find almost all observations along the main diagonal (Type 1 and Type 4). In the empirical reality, however, we find that 52 percent of the observations fall along the opposite diagonal. Out of all the observations, 22 percent are of Type 2, where a negatively biased conception of old people is combined with a pro-old attitude, and 30 percent are of   One way of understanding the four types of attitudinal constellations may be the following: Type 1 means being Consistently Negative in the sense that a negatively biased image of old people is combined with a generally more negative image of what role old people should play in society.
Type 4 means being Consistently Positive in the sense of being quite well informed (or positively biased) about the real conditions for old people, while having a positive attitude toward their participation in society.
Type 2 holds individuals who are inclined to think (a) that older people do not live in their own self-contained houses, have impaired hearing, feel bored and dissatisfied, and seldom see their relatives and friends -at the same time as these type 2 individuals think that (b) there should be more older people among decision makers, that they are underrepre sented in politics, are under-utilized and should be given more authority in order to create a better country to live in. There might seem to be a contradiction between the (a) and (b) sets of opinions, but it could make sense if we borrow and adopt some ideas from the Norwegian philosopher, Harald Ofstad (1972) and from his thought-provoking book Our Contempt for Weakness. Ofstad argues that the different types of conflicting values in society may produce a tendency to look down with condescending pity on those who do not measure up to society's values of productivity, efficiency and independence. But the value patterns of our society also include a tradition in which old age and wisdom are held in high esteem. This tradition is in conflict with value patterns that generate contempt for old people. According to Ofstad, we tend to "resolve" this conflict by hiding the contempt, or by changing it so as to unite it with the respect for old people found in, for example, the Hebrew tradition.
What happens is that some of us transform our contempt into condescending pity. We feel sorry for "the poor, feeble, sick and lonely old people". Moreover, our pity for old people forces us to "construct" them in such a way that the correctness of our pity is validated. We produce a negatively biased image of old people at the same time as we think their conditions ought to be improved.
This line of thought allows us to make sense of our findings: negatively biased conceptions (low on the conceptions scale) combined with a wish to make better use of old people in society (high score on the general proold behavioral disposition scale).
We might call this Type 2 attitudinal constellation Pitying Positive, which, following Ofstad's line of thought, could be related to a hidden or disguised contempt for weakness. This Type 2 attitudinal constellation is also reminiscent of Kalish's (1979) concept of "The New Ageism," which alludes to the patronizing manner in which some care providers and researchers define reality on behalf of old people and decide what they are in need of -particularly by identifying problems and disadvantages. It must be emphasized that the above interpretation is creatively suggestive rather than empirically proved and that other types of qualitative data are needed to confirm the above interpretation.
The Type 3 constellation also appears as a contradiction involving being quite well informed (or having positively biased images of old people), but not being positive on the general pro-old scale. This contradiction too may make sense if we interpret it as follows: Because those who belong to the Type 3 con-stellation have no negatively biased images of old people, there is no need for the fuzziness of the pitying positiveor, expressed another way: "Why coddle the well-to-do elderly?" We call this constellation No Fuzz. Another interpretation of the inner logic of this type is that there is a feeling of rivalry or competition with these well-to-do elderly. With such an interpretation, type 3 might alternatively be labelled 'Jealously negative'. As with the type 2, some qualitative studies are needed for the further exploration of this group.
Looking at each of the above-described attitudinal constellation types, they all make sense in their own way, thus rendering intelligible the contradictions of Type 2 and Type 3. In both of the contradictory categories, we have utilized the new concept Pitying Positive as a key to our understanding. In Type 2, being Pitying Positive is construed as a means for handling the conflict between con-tempt for weakness and a need to honor old age. In the case of Type 3, the positive conceptions do not allow a Pitying Positive solution, paving the way for the No Fuzz position instead.
Who Belong to the Four Types in the Typology?
As a first step in finding out which respondents express these four different attitudinal types, we have checked a number of background variables against these four attitudinal types. Table 4 gives a summary of where we found statistically significant differences.
The table shows that education, age, gender and employment/ unemployment seem to be related to the attitudinal type the respondent belongs to. Being younger as well as being male seems to increase the probability of belonging to the Consistently Negative constellation (Type 1), whereas being older as well as being unemployed seems to be connected with belonging to the Consistently Positive (Type 4). The interesting Type 2 constellation, Pitying Positive, seems to be defined by education and gender. Those with a low level of education as well as women are over-represented in this type. In the equally interesting Type 3, the No Fuzz group, younger respondents as well as men and the employed are over-represented.
In order to more precisely describe how the above attitudinal types are inhabited by different respondents, we have analyzed each type using a CHAID analysis, 23 which, in contrast to, e.g., log-linear models, 23 CHAID (Chi-squared Automatic Interaction Detection) is an algorithm with which a dataset is broken down into sub-categories according to the explanatory power a set of predictors has with regard to a dependent variable. The dataset is subdivided as long as any predictor can produce a statistically significant division, usually with the threshold set at p < .05. (Kass, 1980;AnswerTree, 1998). gives an exact description of the subgroups in which a certain quality (e.g., belonging to a certain attitudinal type) is high (or low). The loglinear models, as well as ordinary regression analysis, only produce mathematical abstractions of the same. Figure 1 shows the outcome of a CHAID analysis in which the Consistently Negative has been the target of analysis. Predictors in the analysis have been the variables shown to correlate with the attitudinal types, i.e., education, age, gender and being employed/unemployed.
As in Table 4, Figure 1 shows the total percentage of Type 1 respondents to be 12 percent. A maximum of Consistently Negative respondents is found among the 158 men in the age category 15-29. In this subgroup, 26 percent of the respondents belong to Type 1, that is, they have both negatively biased conceptions of old people and score low on the general pro-old scale. What we see here is an age-gender interaction such that being a younger male increases the probability of belonging to the Consistently Negative attitudinal constellation.
In order to find the lowest proportion of respondents with Type 1 characteristics, we only have to look at the oldest respondents, among whom only 3 percent belong to the Consistently Negative.
If we for some reason, e.g. change in attitudes, wish to target the specific group in which the Consistently Negative attitudinal constellation is particularly prevalent, the CHAID analysis gives precise information on the make-up of this group. If we, for theoretical reasons, wish to explore the formation of this particular attitudinal constellation, the data suggest we should consider age/gender interactions. There seems to be something in society that produces a tendency toward this consistently negative attitude among younger men.
If we focus on the Pitying Positive attitudinal constellation, in which negatively biased conceptions of old people are combined with a high pro-old attitude, the CHAID analysis in part leads us to other conclusions than those drawn from  In the simple kind of bivariate analysis shown in Table 4, age shows no statistically significant correlation with the presence of the Pitying Positive attitudinal constellation (Type 2). In the CHAID analysis, however, the age variable regains significant explanatory power when the focus is on men only. It is in the sub-category of men between 15 and 49 years that we find the lowest percentage (14%) of respondents with the Pitying Positive attitudinal constellation. Again the CHAID analysis directs our attention to a gender/age interaction. In order to identify the largest proportion of respondents with the Pitying Positive attitudinal constellation, the CHAID analysis simply points to women, who have a larger proportion of Type 2 respondents than do men. Combining the Total 22% n=1673 Women 26% n=810 Men 18% n=863 Age 50-85 22% n=422 Age 15-49 14% n=441 analysis of the Type 1 and Type 2 attitudinal constellations above, we begin to see how attitudes toward old people are gendered. While women, regardless of age, have a somewhat increased probability to be Pitying Positive, young men have an increased tendency to be Consistently Negative.
Turning to the CHAID analysis of the Type 3 attitude constellation, No Fuzz, we once again find that age and gender seem to interact, this time together with the working life situation.
The largest proportion of No Fuzz respondents -those who have good knowledge (or positively biased conceptions) of old people, but at the same time score low on the general pro-old scale -is found among "younger " (age 15-49) employed men. In this group, 43 percent belong to the No Fuzz constellation. When we consider this particular group as well as what this Type 3 attitude constellation stands for, we may again suggest that these men in fact express something like "why coddle the wellto-do elderly?" or "No Fuzz". Quite rightly they know, or think they know, that most old people have a rather good situation in life, and by knowing this they may be inclined to be averse to the suggestions made in the general pro-old scale. But why should this apply in particular to employed men under 50? Could this be a manifestation of a latent Why coddle the well-to-do Figure 3. The No Fuzz generational conflict, in which "younger" employed men are inclined to express aspects of a struggle for power and resources in relation to older people? We shall leave it to others to suggest an explanation for why this does not apply to young employed women. Drawing on the suggestion that the No Fuzz attitude constellation expresses something like "Why coddle the well-to-do elderly?", it makes sense that the lowest proportion of respondents with this orientation is found among the "older" (50-85) unemployed, for whom it might be closer at hand to identify with a group of imagined "poor elderly" and thus easier to fall into another attitude constellation type.
As can be seen in Figure 4, the highest proportion of Consistently Positive (Type 4) is defined by age alone. Simply enough, if you are older yourself, you may have a disposition to be positive toward the advantages your own age group may offer society.
If, however, you belong to the 50-75 age category, your status on the labor market also matters. Being unemployed increases the probability of  being consistently positive. In this latter category, the percentage of consistently positive respondents is almost as high as in the "top category"those between 76 and 85 years. Also in this case, we may suspect that some kind of positive identification process is at hand. The 65+ parliament membership?
As an example of how we could use the new ageism typology, we have targeted the afore-mentioned specific behavioral disposition referred to as "parliament ageism". Figure 5 shows the result of a CHAID analysis in which the proportion of "parliament ageists" is the target of analysis and the attitudinal constellation typology is introduced as a predictor together with education, age, gender and position on the labor market. "No one in parliament should be 65+"

Figure 5. Percentage of "Parliament Ageists"
The analysis shows that the attitudinal constellation typology is what first emerges in our search for subgroups with differing proportions of respondents who believe it to be a good or very good idea that no one in the parliament be older than 65 years of age. Among the respondents belonging to the Consistently Negative attitudinal constellation, the proportion of "parliament ageists" is 47 percent. In the group of respondents who not only belong to this Type 1 category, but also have a low education level, as many as 57 percent are positive toward this age limit. When searching for the subgroup in which the above-mentioned ageist opinion is lowest, the CHAID analysis first collapses the Type 2 and Type 4 attitudinal constellations, where the percentage of "parliament ageists" is more or less the same -20 percent. As regards not being a "parliament ageist" it does not matter whether the high pro-old attitude is combined with negatively or positively biased conceptions.
The division of the above group continues and ends with the subgroup of Type 2+4, more highly educated, where the proportion of "parliament ageists" is 11 percent.
In order to begin to understand the relative importance of the attitudinal type and education, we have run a MCA-ANOVA analysis (not presented), which, in principle, produces an additive model, with all its shortcomings. This run produces a beta value of 0.23 for the attitudinal type variable and 0.20 for the education variable. Thus, within such an additive model, attitudinal type and education seem equally important. That is, regardless of educational level, position in the attitudinal typology is of importance, and regardless of position in the attitudinal typology, educational level is of equal importance.

Discussion
By referring to our own previous work as well as that of other more contemporary scholars, we have argued that ageism may be fruitfully understood and studied as a constellation of dimensions borrowed from the traditional concept of attitudes. In doing this, we have suggested a new typology. There is certainly a need to cross-check this typology and the inner meaning of its types, by means of qualitative studies of individuals belonging to each of the types. There is always the possibility that the inner meaning of a type will diverge from the interpretations we have made. This is a need that the present typology shares with all other theoretical and conceptual novelties. It is also the task of future research to connect the typology with the emotional dimension of attitudes/ageism, which we, lacking proper data, have not been able to address. With these remarks in mind, the following discussion may be justified.
When both the cognitive and the behavioral disposition components are considered at the same time, we find attitudinal constellations that at first glance seem contradictory, but that, when considered individually, have their own logic.
In particular we have identified the two constellations the Pitying Positive and the No Fuzz.
The Pitying Positive are those who have a negatively biased conception of old people at the same time as they seem prepared to act to give old people a greater role to play in society -thus combining negatively biased conceptions with a pro-old behavioral disposition. Empirically, nearly a quarter of Swedish respondents between 18 and 85 years of age belong to this category. This should alert researchers to consciously include this type of attitudinal constellation whenever studying ageism.
Also, the empirical analysis has shown that women are more likely to belong to the Pitying Positive. Men, on the other hand, are more likely to belong to the other opposite constellation -the No Fuzz.
The No Fuzz group, comprising 30 percent of the Swedish respondents, are those who combine positive conceptions of old people (no negative bias) with little or no positive pro-old behavioral dispositions. This can be understood as the mirror image of the Pitying Positive, such that positive conceptions are combined in an attitude constellation expressed as "Why coddle the well-to-do elderly?" or "No fuzz". The presence of this constellation is as high as 43 percent among younger employed men, signalling an age/gender/employment interaction, which may be worth exploring in future research on ageism.
Together the No Fuzz and the Pitying Positive constellations add up to no less than 52 percent of the attitudinal constellations, which again points to the need to distinguish and include these kinds of "contradictory" constellations in future research on ageism. By doing so, we may, as shown above, obtain a fuller understanding of why and how different groups of individuals include as high as 57 percent or as low as 11 percent "parliament ageists", who consider it a good or very good suggestion that no one in parliament be over the age of 65.
Our results also suggest the need to apply an age/gender perspective when studying ageism, as several of the ageist positions seem to be characterized by age/gender interactions.
With respect to our understanding of very specific behavioral dispositions, as "Parliament Ageism" in our example, it may be noted that the absence of such ageism may sometimes arise for "the wrong reasons". Being a Pitying Positive decreases the probability of being a "Parliament ageist". A question for future research concerns what the outcomes of information campaigns actually are. It has mostly been taken for granted that proper knowledge about old people is beneficial for attitudes and ageism, but the present research reveals the possibility of other outcomes. With good knowledge, the probability of being Pitying Positive decreases, and instead the probability of becoming either Consistently Positive or No Fuzz increases. At this point, we know nothing about the degree to which increased information could increase the proportion of No Fuzz people, who, as mentioned above, are negative toward parliament membership for those above 65 years. A hypothesis derived from the above data is that more correct (or positively biased) information produces an increase in the proportion of No Fuzz young men, and perhaps also an increase in the proportion of Consistently Positive women, thus creating something of a polarization between men and woman. This is, however, merely a hypothesis, to be targeted in future research on how ageism may be both gendered and age-related and on the interaction between these two factors. As hinted at before, it is also a task for future research to explore which kinds of emotions are related with the various types. A preliminary hypothesis is that basically positive emotions go with the Pitying Positive and the Consistently Positive, while the No Fuzz and the Consistently Negative are connected with neutral or negative emotions.