Reproduced from 1988 Conference Proceedings, Pp
Attitudes toward adult education in Britain and the United States
Reproduced from 1988 Conference Proceedings, pp. 131-135 SCUTREA 1997
Attitudes toward adult education in Britain and the United States
Gordon G. Darkenwald, Rutqers University
A study of attitudes toward adult education conducted in the U.S. was replicated in Britain. The objectives of the replication were to explore the extent to which American and British attitudes were congruent, predicted participation-related behaviour, and exhibited similar correlations with respondent background characteristics. The U.S. and British findings were comparable in certain respects, but divergent in others.
A typical definition of attitude is ‘an internal state which affects an individual’s choice of action toward some object, person, or event’1. It follows, of course, that attitudes toward adult education significantly influence the likelihood of participation in adult education. Although the construct of attitude is central to theories of participation behaviour2, only a handful of attempts have been made to measure attitudes - all of them unsuccessful. Attitudes are important not only because they affect participation in adult education, but also because they determine in part public support for the provision of adult education in the community and workplace.
Theory of attitude measurement
Due in part to a lack of conceptual and technical sophistication in scale construction, attitude measures are often poor predictors of behaviour. Rokeach3 has cogently addressed the former deficiency in his formulation of the two-attitudes theory. Briefly, Rokeach postulates that behaviour is affected by two types of attitudes: attitude toward object e.g., adult education) and attitude toward situation (e.g., participating in 3 discussion group). By itself, attitude toward object has not exhibited consistent relationships with behaviour. Attitude toward the situation must also be measured within the same scale if the A-B correspondence is to be maximised.
The Adult Attitudes Toward Continuing Education Scale (AACES) was developed and pilot-tested over a six-month period by a team of 10 Rutgers doctoral students supervised by the present author. A detailed description of the procedures employed to construct AACES can be found in Darkenwald and Hayes4 .
The final form of AACES consists of 22 items on a Likert-type five-point scale. Seven of the 22 measure attitude to situation (e.g., ‘Being in a classroom makes me feel uncomfortable’) and the remainder attitude to object (e.g., ‘the need for education continues throughout one's lifetime’). The version of AACES administered in Britain was identical to the original except for necessary Anglicisation of the directions and a few words such as ‘program’ (programme). The reliability of AACES was high and identical for both the U.S. and Britain samples (alpha = .90). Evidence of predictive and concurrent validity is presented in the findings section of this paper.
The directions to respondents were equivalently worded in Britain and the United States. Adult education was defined very broadly to include any subject, purpose, or format. The British directions read as follows.
Adult education is defined here as any education provision for adults (including job training) organised by local adult education centres, further education/technical colleges, community groups and clubs, employers, universities, unions, churches, associations, hospitals, etc. It includes all ways of learning and all subjects that adults wish or need to study.
Please read the following list of statements. Each expresses an opinion about adult education. There are, of course, no right or wrong opinions. For each statement, circle the number that best describes your feelings about it: 1=strongly disagree, 2=disagree, 3=undecided, 4=agree, 5=strongly agree. Should you change your mind, cross-out your first answer.'
A four-item ‘yes/no’ Behavioural Index was developed to assess attitude-behaviour (A-B) correspondence. Multi-item measures are preferred for this purpose in that ‘a behavioural score based on conceptually related and intercorrelated behaviours should be more reliable than a single behaviour, and this in turn should increase the A-B relationship’5. The first three items inquired about correct participation in adult education, frequency of participation, and past participation; the last asked respondents ‘Have you ever suggested to another adult that he or she engage in some kind of adult education or training’? The alpha reliability coefficient for the Index was an acceptable .63.
Sampling and Data Collection
The U.S. data were obtained from 275 adults (person 18 or older and not full time students) resident in central New Jersey. Since lack of funds precluded drawing a random sample, purposive sampling aimed at securing responses from a cross-section of the general public was employed. The data were collected personally by the members of the research team. Although a strictly random sample would have been preferable, it was not essential for the purposes of the research. To explore and compare relationships among variables (rather than estimate population parameters) requires only that the number of cases be sufficiently large and heterogeneous to permit sophisticated statistical analyses. Ultimately, external validity, as is always the case, can only be established by multiple replications.
In Britain, a market research firm was engaged to draw a random sample or the adult public from the electoral lists of England. (NB: this sample was not the same one employed for my other conference paper). Wales and Scotland were excluded because of their cultural and linguistic distinctiveness. Six hundred questionnaires were mailed with cover letters and stamped, pre-addressed return envelopes. Follow-ups of non-respondents were conducted at three-week intervals. Two hundred fifty-one useable questionnaire were returned. Adjusted for postal service ‘undeliverables’ (n-115) and illness, death, disability, etc. (as indicated by letters or notes with returned blank questionnaires - a total of 34), but not refusals (n-19), the return rate was a respectable 55.6%. Bias estimation procedures revealed that the U.S. sample was over-represented by adults with high incomes and high educational attainment. The British sample, however, was quite close to being representative of the general adult public.
Description of Respondents
Only a cursory overview of the demographics for the two samples can be given here. For England, the basic statistics are these. Sex: 55.4% female; Age: mean=44 years; Employment status: 42.3% employed full time; Educational attainment: 40.2%, no formal credentials; 33.1% CSE to GCE A-level secondary school credential; 7.2% two or three year post-secondary school credential (e.g., ONC, HND); 6.4% university first degree; 8% other, including postgraduate study or degree; annual gross household income: 19% less than £5,000; 27.1% £5,000 to o£10,000; 31.2% £11,000 to £20,000; 9.3% £21,000 to £29,000; 6.1% £30,000 or more; and 6.9% (mainly women) 'don't know'. In May, 1987, the time of the survey, one pound sterling equalled $1.60. The New Jersey statistics indicate a pronounced discrepancy in income and educational attainment, which simply mirrors the reality of the two countries. However, at secondary level and below one should not draw a simple conclusion since British school completers who do not sit for exams (the majority) receive no credential.
The New Jersey demographics are these. Sex: 53.5% female; age: mean=39 years; Employment status: information not obtained; Educational attainment: 5.2% no credential; 27.5% secondary school credential; 26.8% post-secondary study or credentials; 21.2% college/university first degree; 19.3% postgraduate study or degree; Annual gross household income: 7.6% less than $15,000; 32.8% $15,000 to $29,000; 32.8% $30,000 to $44,000; 26.7% $45,000 or more. With respect to race/ethnicity, 14.5% of the U.S. respondents described themselves as black; only 3.6% of the English sample indicated minority group status, mostly Indian and West Indian.
Describe statistics were calculate for the AACES items and total scale scores for both samples. Negatively worded items were reverse-coded; thus the higher one's score, the more positive one’s attitude toward adult education. Since the item and total scale statistics differed little, sophisticated statistics, such as the Mann-Whitney U test, were not calculated. The principal research objective - determining the relationship of attitudes (and selected demographics) to participation behaviour - was addressed by computing least square multiple regress equations for both the British and American samples. The independent variables in the regression equations were attitude score and sex, age and educational attainment (the only demographics available for both countries). The dependent variable was the Behavioural Index. Model-fitting statistics revealed that the residuals for U.S. educational attainment were negatively skewed. Consequently, a square transformation of the dependent variable was required to produce an appropriate model. Correlational analyses were utilised to explore the relationship between respondent demographics and attitude scores. The criterion for statistical significance was p<.05.
The results of the multiple regression analyses are presented in Tables 1 and 2. In both equations, sex was insignificant when controlled on the other independent variables and therefore deleted from the summary tables. For both countries Attitude (positive) and Educational Attainment (higher) were the principal predictors of participation-related behaviour. However, it is noteworthy that in Britain the influence of attitudes, best indicated by BETA (a sort of purified correlation coefficient), was substantially greater than in the USA. This could be due to the higher levels of educational attainment in the USA. Prior research has established that this variable is by far the most potent predictor of adult education participation.
Oddly, the total ‘explained variance’ (R2) was identical for the two countries. Although R 2 I S on the order of .28 are quite respectable in behavioural research, it should be kept in mind that more than two-thirds of whatever accounts for participation behaviour remains unidentified. Future research must take into account other promising variables, such as deterrents and motivation.
Why age (being somewhat older) should exhibit a modest but statistically significant effect in Britain but not in the U.S. is puzzling. Perhaps the explanation is related to the fact that middle-aged and older women are the main participants in non-vocational adult education in Britain6.
Table 1. Summary of Multiple Regression Analysis: BritainIndependent Variable / Simple r / Multiple R / R2 / BETA
Attitude / .50** / 50** / .25 / .48**
Education / .26** / .52 / .27# / .18*
Age / .01 / .53 / .28# / .13*
*P.= .02; **P.= .0001; #=increment in RZ sig. at P.= .02
Table 2. Summary of Multiple Regression Analysis: USAIndependent
Variable / Simple r / Multiple R / R2 / BETA
Education / .47* / .47 / .22 / .47*
Attitude / .35* / .53 / .28 / .26*
*P.= .0001; #=increment in R2 significant at P.= .0001
Some statistically significant correlations were found between AACES scores for both countries and respondent socio-demographics. In both countries more positive attitudes were associated with higher levels of education. In the USA, sex (female) was correlated with positive attitudes and so too was income. In Britain, younger adults exhibited more positive attitudes toward adult education; in the U.S., no relationship was observed. The explanations for most of the observed discrepancies are not difficult in light of the findings presented above.
It appears that attitudes toward adult education are reasonably comparable in Britain and the United States. Attitudes in both countries were generally positive, and the low- and high-ranked AACES items were nearly identical. However, there were notable discrepancies between Britain and U.S. with regard to the relationships between attitude scores and respondent demographic characteristics: plausible explanations can be preferred, but could well be wrong. Similarly, the multiple regression analyses, although accounting for the same variance in participation behaviour, exhibited differences in the predictive power of educational attainment and attitude toward adult education. The relatively low variance in level of formal education in Britain probably lies at the root of the discrepancies.
Anderson, R. and Darkenwald, G.G. (1979) Participation and persistence in American adult education. New York: The College Board
Hui, C. and Triandis, C. (1985) Measurement in cross cultural psychology. In Journal of Cross-cultural Psychology, 16, 131-152
Gagne and Briggs, 1974, p.62
Cross, K.P. (1981) Adults as learners. San Franciso: Jossey-Bass; Darkenwald, G.G. and Merriam, S. (1982) Adult education: foundations of practice. New York: Harper & Row; Houle, C. O. (1985) Structural features and policies promoting and inhibiting adult learning. In Knoll. J. Motivation for adult education Paris: Saur, 64-73; Rubenson, K. (1977) Participation in recurrent education. Paris: OECD
Rokeach, M. (1968). Beliefs, attitudes, and values. San Francisco: Jossey Bass
Darkenwald, G.G. and Hayes, E. (1988) Assessment of adult attitudes toward continuing education. In International Journal of Lifelong Education, 7, in press
Schuman, H. and Johnson, M. (1976) Attitudes and behaviour. In Annual Review of Psychology, vol. 36. New York: Annual Reviews; p. 183
McGivney, V. (1988). Personal communication