The article will be available at http://www.aidsmap.com/cms1061207.asp after a few weeks.

I think it can be generalized as a good method for reviewing all kind literatures. To illustrate this method, take their study as an example:

First, They indentified seven themes when they were reviewing 268 studies :

Theme 1: Young people subjectively assess the risks from sexual partners on the basis of whether they are “clean” or “unclean”;

Theme 2: Sexual partners have an important influence

on behaviour in general;

Theme 3: Condoms can be stigmatising and associated with lack of trust;

Theme 4: Gender stereotypes are crucial in determining social expectations and

behaviour;

Theme 5: There are penalties and rewards for sex from wider society;

Theme 6: Reputations and social displays of sexual activity or inactivity are important;

Theme 7:Social expectations hamper communication about sex.

They did 3 steps to complete this categorizing :

Step one, generate theme independently reviewing all studies;

Step two , refine these dozens of codes through discussion and the use of constant comparison within and between codes to ensure that they accurately reflected the material.

Step three, identify correlations between the different themes, grouping them into the broad overall themes. They got the seven themes above.

Second, they coded all 268 studies with 1, 2, 3, 4, 5, 6 and 7 according to the content of each study.

Finally, studies were classed as primary—of high quality or containing empirical data about sex (ie, specific reports about sexual events rather than about attitudes or opinions), or both, or they were classed as secondary—lower quality, with no empirical data about sex.

After complete the above, you can present the results in a table as following:

Britten N, Campbell R, Pope C, Donovan J, Morgan M, Pill R (2002) Using meta ethnography to synthesise qualitative research: A worked example. Journal of Health Services Research and Policy 7:209-215

Greenhalgh T, Robert G, Macfarlane F, Bate P, Kyriakidou O, Peacock R (2005) Storylines of research in diffusion of innovation: a meta-narrative approach to systematic review. Social Science & Medicine 61:417-430

Marston C, King E (2006) Factors that shape young people’s sexual behaviour: a systematic review. The Lancet 368:1581-1586

At the beginning, socially desirable response reflected only a response style (Pauls and Stemmler 2003). However, soon it emerged that social desirability could be conceptualised as a personality style (D.P. Crowne and Marlowe 1964). Recall Raymond in “Everybody loves Raymond”. He loved being called ‘a nice man’ and brought yummy snack to clubs always even when his wife told him it’s no so fair if other didn’t do so. Or Bree in Desperate Housewives who wants to be perfect all the time desperately.

Social desirability can cause bias in research, especially in psychological and medical research. Social desirability often manifests in the overestimation of desirable traits or behaviors and the underestimation of undesirable traits or behaviors (Paulus 1991).

For example, there is significant comorbidity between alcohol abuse and anxiety disorders. However, the prevalence of anxiety disorders in alcoholic samples has been found to be higher than the occurrence of alcohol abuse in anxiety disorder samples. (Cox et al. 1994). One of possible reasons is alcoholism is less socially acceptable than anxiety disorders. People also intend to overreport their physical activities, particularly when they know investigators are experts in health since physical activity has been considered to be a socially desirable behavior (Motl et al. 2005).

Social desirability bias also can affect research in sexual behavior. If investigators are sexual health workers other than column writer from Cosmopolitan, it’s very likely people will underreport the number of their sexual partners and overreport their protected sexual intercourse, such as condom use. Well, the problem here is how likely it is.

In my study, sex workers may overreport their use of male condoms. There is no existed evidence so far as I know in the UK to prove that they did so definitely. I am thinking there are several possible solutions to this issue. ……

Cox BJ, Swinson RP, Direnfeld DM, Bourdeau D (1994) Social desirability and self-reports of alcohol abuse in anxiety disorder patients. Behaviour Research and Therapy 32:175-178

D.P. Crowne, Marlowe D (1964) The approval motive. Wiley, New York

Motl RW, McAuley E, DiStefano C (2005) Is social desirability associated with self-reported physical activity? Preventive Medicine 40:735-739

Pauls CA, Stemmler G (2003) Substance and bias in social desirability responding. Personality and Individual Differences 35:263-275

Paulus DL (1991) Measurement and control of response bias. In: J.P. Robinson PRSaLSW (ed) Measures of personality and social psychological attitudes. Academic Press, San Diego (CA) pp 17-59

and some presented results from NSSAL 2000