Becker et al 2002 (contemporary)

• Why Fiji

  Has extremely low prevalence of eating disorders - only one reported case in mid 1990s

  Nadroga province of Fiji was selected - lack of exposure to TV until mid 1995

  Fijian culture encourages robust appetites and larger figures
• Aim

  To look at impact of introduction of TV on eating behaviour and attitudes in Fijian adolescent girls
• Procedure (1)

  Cross sectional design used for girls from secondary school

  First group tested in 1995 - 63 girls w mean age of 17.3 years

  Second group tested in 1998 -65 girls of same age

  No sig diff of girls in terms of mean age or body mass index
• Procedure (2)

  Girls asked to complete a 26 item eating attitude test (EAT-26) (questionnaire)

  Girls who scored over the threshold for dysfunctional eating based on EAT-26 were asked to respond to semi-structured interviews to confirm symptoms
• Findings

  Weight did not differ between two groups of girls with BMIs of 24.5 and 25.9 respectively.

  TV viewing did increase with 41% having a TV at home in 1995 and 71% in 1998

  Dysfunctional eating increased w 29.2% of girls scoring over threshold of 20 on EAT-26 in 1998 (12.7% in 1995)

  11.3% reported purging in 1998 (none in 1995)

  1998, 74% said they felt too large or fat and 69% had dieted - alien to Fijian culture

  Interviews suggested girls considered TV characters to be role models
• Conclusion

  Study suggests TV can have a negative impact on eating attitudes and behaviours
• Strength 1

  Reliable measurements - studies have generally shown good reliability for EAT-26, the main measure of eating behaviour in this study

  Eg Juan Rivas et al found good internal reliability - correlation of +9 for each item and overall result

  SO likely that diffs between groups were due to real changes over time
• Strength 2

  Application to treatment or avoidance - study show modelling in mass media can have negative impacts on body image and eating behaviour

  Can be applied to advice for TV programming and advertising - understanding value of 'larger size' models

  Also, if these behaviours can be learned through modelling, can be unlearned in similar way
• Weakness 1

  Poor generalisability - unrepresentative sample so can't generalise findings as Fiji traditionally has distinctive attitudes to body type, eg high BMIs are considered attractive

  Clash between Fijian cultural norms and these portrayed on American TV were much greater than what most countries would be

  SO results may have been affected by culture and could tell us little about affect of TV on body image in other countries
• Weakness 2

  Issues w validity eg, we would expect that scores on EAT-26 would predict eating disorders - but was not case w Brazilian women (Nunes et al)