Category: Sociological concepts

Experiments are the only method educational researchers can use if they wish observe the effects of one specific variable on student behaviour or outcomes (results).

Experiments are probably conducted more by schools themselves to test out things like new teaching techniques before rolling them out to the whole school, and there are also several examples of policy changes providing us with some examples of ‘natural experiments’, as when academies were introduced, they allowed researchers to compare the performance with LEA schools.

Examples of sociologists going into schools to conduct their own research are a lot rarer, and laboratory experiments on how social factors relate to educational performance are rarer still.

This post provides examples of all ‘four types’ of these experiment. It has been written primarily for students of A-level sociology studying the Methods in Context aspect of the specification.

Field Experiments within Schools by schools themselves

There are a number of variables schools might try to change in order to improve student behaviour, performance, or just to enhance student well-being.

Experimenting with setting and streaming, the gender mix of classrooms, different teaching techniques, online learning, or even the length of the lessons themselves are all possible focuses for small scale experiments.

I discuss this more in this post: experiments within schools.

One of the most extreme field experiments conducted recently was by a school in Devon, in which they subjected some of their students to a Chinese style of teaching, involving Chinese teachers, for a three month period. For more on this, please see this post.

Field Experiments by Sociologists within Schools

The classic field experiment relevant to education is Rosenthal and Jacobsnen’s Pygmalion in the Classroom, in which they set out to measure the impact of high teacher expectation on student performance.

They went into a school, and tested a sample of the students, keeping the actual results hidden from the teachers. They then told the teachers that a randomly selected sample of students were especially gifted (when in reality the students had a range of abilities).

The researchers then left the school, returned some months later and re-tested all the students. They found that the ones who teachers had been told were higher ability had improved at a faster rate than the rest.

The conclusion is that this supports the Self Fulfilling Prophecy Theory, however other repeat experiments have yielded different results.

For more details on this experiment, please see this post

Natural Experiments and Education

There have been two notable government policies which have introduced new school types in recent years: Academies and Free Schools.

We now have several years of data to compare the performance of both of these types of school with regular Local Education Authority schools, which is a natural experiment.

We could also do the same at a global level, by looking at the PISA league tables and then looking at what features the education systems of the top performing countries have in common, if any.

Coronavirus has also provided us with an interesting opportunity to measure the effects of online learning on education. Some recent initial studies report that poorer students are negatively impacted more than wealthier students.

Be careful discussing ‘natural experiments’ in an exam, as we are getting into ‘secondary data’ here rather than pure experiments, but there are links!

Laboratory Experiments relevant to education

There are a couple of interesting historical examples:

Charkin et al (1975) conducted research with a sample of 48 university studetns who each taught a lesson to a 10 year old boy.

One third of the university students were told they boy was highly motivated and intelligent

One third were told he was poorly motivated and with a low IQ

One third were given no information

Charkin et al videod the lessons and found that those in the high expectancy group made more eye contact and used more encouraging body language than the low expectancy group.

This seems to suggest support for labelling theory.

Mason (1973) looked at whether negative or positive expectations had a greater effect.

Teachers were given positive, negative or neutral reports on a pupil. The teachers then observerd video recordings of the pupil taking a test, watching to see if any errors were made, and then asked to predict the pupil’s end of year attainment.

Mason found that the negative reports had a much greater impact on the teachers’ expectations than the positive reports.

A much more recent experiment, aired by the BBC, showed how simply having a mobile phone on a desk lowers the test scores of students. For more details on this, please see this post.

(Source Webb et Al A-level Sociology, Book 1)

Please click here to return to the homepage – ReviseSociology.com

Social Surveys are probably the most practical method researchers can use to research education.

Both teachers and students will be relatively used to filling in questionnaires as part of providing feedback to improve lessons and school procedures, and they are relatively quick to complete compared to interviews, so they should cause relatively little disruption to the school day.

Practical Factors

Surveys are a good choice of method if you wish to collect data from large samples in a short space of time, and because students (when they’re in class) are a ‘captive audience’ this could make this a very time efficient method for researching in schools.

If a researcher can gain access and consent they could, with the co-operation of a head teacher, get hundreds of students to complete the same questionnaire in multiple class in one day, or on smaller scale, a researcher could work with a few teachers to get a slightly smaller sample.

If the researcher puts a survey online, management might give teachers more flexibility when they get their students to complete the survey – say in one tutorial session over a two-week period.

A closed-question questionnaire might be a good method for researching teachers, given their tie constraints.

Parents would be the most difficult group to research using the questionnaire method, as they spend less time in-school, so gaining access to them would be difficult, this would probably have to be done via the school, who could direct parents to questionnaires via newsletters, or may parents evenings could be used by the researcher to administer surveys.

Gaining access

Because surveys are relatively quick for respondents to complete and it’s obvious from the outset what questions are being asked (which wouldn’t be the case with interviews), then it should be easy to convince schools to gain access to students, teachers or parents, compared to more intensive qualitative methods.

Theoretical Factors – Representativeness, validity and reliability.

Representativeness

Schools have ready made lists of students, which would include details of their gender, ethnicity, FSM and SEN status, address (as a proxy of broader class status) and prior educational achievement.

IF a researcher could thus gain access to such a sampling frame, it would be very easy for them to get a representative sample of different students, or to select only one type of student (all boys for example), depending on the purpose of their research.

However, getting access to such a list with all of the above details may not be possible because of GDPR (data protection) issues, unless researchers work with school staff who select a representative sample on their behalf.

Response rate

Biased samples might mean a low response rate for some types of respondent

Questionnaire research might suffer from selection bias – pro-school pupils are much more likely to take them seriously, but more rebellious students who do not like authority might either not fill in a questionnaire or deliberately lie out of spite against the system.

Working-class parents might be less willing to fill in a questionnaire truthfully about their parenting practices, whereas for middle-class parents this would be more a positive affirmation of their ‘good parenting’

Non-native English speakers might not be able to understand the questions if the questionnaire is not in English. Although today there might well be programmes online that can translate online questionnaires.

Validity

Because questions are written in advance, this does not allow for an in-depth exploration of respondents’ thoughts and feelings, hence validity may be limited for some topics.

Researchers also must be careful that concepts (such as cultural capital) are operationalised in such a way that children (especially young children) can understand them.

As mentioned above, the formal nature of questionnaires may not yield valid data from rebellious students – and the more formal and more test-like the conditions of completing a questionnaire, the more likely this is to be the case.

Reliability and making comparisons

Questionnaires do allow for excellent reliability, which is useful if findings are to be used to inform educational policy – it allows the research to be scaled up and generalised to more areas easily.

This is also a good method for exploring differences between students from different social class, ethnic backgrounds, as well as gender differences, which is a huge topic in the sociology of education  

Ethical issues and questionnaires

A big strength of questionnaires is that it is easy to make them anonymous and so to keep pupil, teacher and parent data confidential, so they’re good for exploring sensitive topics.

Cross National Comparisons involve researching a specific social institution, trend, or phenomenon in two or more countries using the same research methods, with the intention of comparing how this institution, trend, or phenomenon manifests in different socio-cultural settings.

Researchers might choose to focus on broad topics such as the education system, or a specific trend such as the suicide rate, and they may use analysis of already existing secondary data to do this, or conduct their own original primary research.

The aim may be to seek new explanations for similarities or differences or to gain a deeper understanding of social reality in different national contexts.

Examples of Cross-National Comparisons

Durkheim’s classic 1897 study of Suicide

Emile Durkheim’s study of Suicide was the first ever study to call itself a work of sociology. Durkheim wanted to find out whether the very personal act of suicide was shaped by social factors.

He used official statistics on a range of factors such as the religion of a country (Protestantism or Catholicism), the rapidity of economic growth, and the divorce rate (among other things) and then correlated these with the suicide rate.

He famously found that countries with lower rates of social integration and social regulation have higher suicide rates.

(NB this is a summary of one aspect of the study, it’s a bit more complex than this!)

He was famously criticised for many reasons, one of which was his failure to take account of the stats lacking validity.

Wilkinson and Picket (2010) The Spirit Level

In 2010 these two researchers looked at income equality in several different countries (the gap between the highest and lowest earners.

They found that a higher level of inequality was associated with all sorts of social and cultural problems such as:

• Higher rates of imprisonment
• Higher levels of obesity
• Higher rates of suicide.

A 2014 study on immigration and gender equality

In 2014 Roder and Muhlau noted that there is considerable cross national variation in attitudes towards gender equality, and they were interested in exploring what happens to the attitudes of immigrants who move from a less gender egalitarian culture to one in which gender equality is more activity promoted.

They started off with the following hypothesis:

(a)Second- generation immigrants have a more egalitarian-gender ideology than the first generation and (b)the gender relations of the origin country exert less influence on the gender attitudes for second generation immigrants than for first generation immigrants.

To test the hypothesis the authors used data from the European Social Survey (2014), which is conducted in several European countries by structured interview every two years.

They defined ‘second-generation immigrant’ as anyone who was born in the country in which they presently lived but had at least one parent who was born abroad.

To measure gender egalitarian attitudes, they relied on two questions from the questionnaire, both being Likhert Scale questions, consisting of statements which people had to agree or disagree with on a 1 to 5 scale.

To measure gender equality in the origin country, researchers relied on indicators such as female representation in parliament, professional jobs, and income differences between men and women.

• ‘When jobs are scarce, men should have more right to do a job than women’
• ‘A woman should be prepared to cut down on her paid work for the sake of her family’

The hypothesis was broadly proved correct

The Limitations of Cross-National Studies  

Cross national studies tend to be very large in scope, and so can require considerable funding to carry out, securing funding can be a problem (an agency based in one country may be reluctant to fund research that takes place in multiple countries, so sources of funding are likely to be limited to international agencies, rather than national governments).

Data collected from official government sources (official statistics) may not be comparable – the categories used and the methods of collecting the data may differ from country to country.

Data will have to be translated, and there is a problem of this translation being insensitive to specific national and cultural contexts. Cross-national research helps to overcome assumptions we might make about life in other countries.  

According to radical feminist theory, sexual violence is one of the main ways in which men control women in relationships, and in society more generally, and one of the main reasons violent men can maintain their control is because there is simply so much silence surrounding the issue – it’s not only that the victims are often silent, but also other people close to them are reluctant to discuss sexual violence, even when they know it is going on.

A recent book: The Anatomy of Silence explores why this is through 26 stories written by women who have been victims of Sexual Violence, and why they were silent about their experience.

The book covers several different examples of sexual violence, from child-abuse, through to rape on college campuses, and includes experiences from men as well as women.

The book seems very radical in its approach (compared to the usual silence around the issue) and criticises the #MeToo movement for highlighting high profile, celebrity cases of sexual violence, rather than the everyday violence experienced by ordinary people, and for failing to address what needs to be done to prevent sexual violence in the future.

The main focus of the book seems to be the problem of silence around domestic violence – with those close to the victims often being the ones to silence them, out of shame, whether it’s their mothers or the university as an institution not wanting to acknowledge the extent of the problem for fear of their reputation.

There is also a focus on both men and women together having to take more responsibility to tackle the culture of ‘toxic masculinity’ and to call out friends, family members and colleagues more for acts of what you might call ‘every day sexism’.

A good example of Feminist research methods

• It is qualitative in-depth research, which gives readers insight into the feelings, the emotional experiences of victims.
• The research comes from the respondents, it is victim-led. NB In the blog post linked above, the editor of the stories is only mentioned briefly, write at the end of the article.
• The research is overtly political, aimed at empowering silenced victims.
• Also note the clever use of language – it’s called ‘Violence’ not ‘domestic abuse’, the later sounds a little softer.