MEDD8001 Chap.7 Quantitative Research and Familial Influences
Quantitative Research and Familial Influences
What families actually transfer
The intuitive answer to why children from higher-status families reach higher attainment is money, and the lecture opens by offering three options: more wealth to spend on learning, a larger network of people with useful resources, or parents who know the education system and have the skills to support a child inside it.
Those are the three capitals in everyday language, and most people pick the first while the research literature gives substantial weight to the other two.
Family capital is defined as whatever a household holds that it can spend or put to work in order to live better and function better, existing in economic, social and cultural forms that produce both educational attainment and the habits and preferences shaping later outcomes.
Three capitals, and three axes
Economic capital is wealth, comprising net worth and financial wealth, with net worth defined as the current value of marketable assets less debt and financial wealth as that figure again once the equity held in the home you live in is taken out.
Social capital has two components, networks that vary by strength of tie, homogeneity and openness, and exchange or support that may be instrumental, symbolic, informational, or the informal introductions that open up people and places a household could not otherwise reach.
Cultural capital is what a person knows, does by habit, believes and values, in so far as it helps them find their way through an institution, existing embodied in the person, objectified in possessions, or institutionalised in recognised qualifications.
The investment model adds three axes: the total volume of capital, its composition, and its stability or change, which together name three different research questions.
Published studies, and one counter-intuitive result
The lecture assembles studies that operationalise these differently: parental involvement related to academic outcomes directly and through school engagement; parental involvement in young children's foreign language learning in Hong Kong; how far a household's standing and a parent's expectations went with results in mathematics and in problem solving, in mainland China; and family influence on subject and university choice.
The most useful result is a sign.
In the mainland China study two components of socioeconomic status, home educational resources and parental education, positively predicted parental expectations, while family wealth predicted them negatively.
A composite labelled socioeconomic status contained one component pulling the opposite way from the other two, which is invisible once the three are summed into a single index.
The machinery: sampling, scales, validity, reliability
Probability sampling comes in four forms, simple random, stratified, cluster and systematic, each defined by its selection rule.
Non-probability sampling covers convenience, purposive and quota sampling. Neither column is the better one; the right-hand column becomes a fault only when a write-up generalises as though the left-hand column had been used, and the lecture answers two standing worries directly by stating that quantitative work requires neither a large sample nor a random one.
Four scales follow, nominal, ordinal, interval and ratio, and the arithmetic you may perform follows from which one you hold.
Validity asks four questions, content, concurrent, predictive and construct, while reliability asks one, whether the measure yields consistent results, with named conditions that raise and lower it.
Reading a quantitative paper the way the course reads one
The lecture walks a published study section by section, and the sequence is the template for a methods chapter: participants reported in enough detail that a reader can picture who was studied; procedure reported so that collection and ethics adherence are visible; measures reported with the source of the instrument, the number of items, a sample item, the response scale, how the score was computed and the reliability obtained; and an analytic strategy that explains how each question will be answered.
One pattern from the results is worth carrying away: a positive path into a mediator can meet a negative path out of it, and a significant first leg does not establish the chain.
What this chapter covers
- 01
Three capitals as three answers to one question
- 02
Net worth and financial wealth, defined
- 03
Networks and exchange: the two components of social capital
- 04
Embodied, objectified and institutionalised cultural capital
- 05
Volume, composition and change: three axes, three questions
- 06
One composite whose components pull in opposite directions
- 07
Four probability and three non-probability sampling rules
- 08
Nominal, ordinal, interval and ratio scales
- 09
Four validities against one reliability
- 10
Observer bias against observer effect
- 11
Central tendency, dispersion and the correlation coefficient
- 12
Participants, procedure, measures, analytic strategy
Four decisions already made, and one not noticed
- 3Name what the sampling choice does and does not license.
- 3Name the scale and the composite issue.
- 4Name the unnoticed decision about whose construct is being measured.
Key terms
- Economic Capital
- Family wealth, comprising net worth, the current value of marketable assets less debt, and financial wealth, which is that same figure once the equity held in the home you live in is taken out.
- Social Capital
- A family's web of connections together with the exchanges running through them, which may be instrumental, symbolic, informational or a matter of the informal introductions that open up places a household could not otherwise reach.
- Cultural Capital
- What a person knows, does by habit, believes and values, in so far as it carries them through an institution, held in embodied, objectified and institutionalised forms.
- Stratified Sampling
- Dividing a population into subgroups and then selecting randomly within each, so that the sample reflects the subgroup structure by design rather than by luck.
- Quota Sampling
- Drawing a non-probability sample so that it approximates known proportions of the population on selected characteristics.
- Ordinal Scale
- A scale that rank-orders cases by how much of a characteristic they possess without equal intervals between points, which is where agreement ratings sit.
- Construct Validity
- Whether a test adequately measures the psychological construct it claims to, as opposed to measuring something correlated with it.
- Internal Reliability
- The consistency of a multi-item measure, reported so that a reader can judge whether the items belong together before the composite is used.
- Observer Effect
- The change in participants' behaviour caused by being observed, which is a problem in the participants and needs a different remedy from observer bias in the researcher.
Quantitative Research and Familial Influences FAQ
Do I need a large sample to do quantitative research?
The lecture puts this to students directly and the answer is no, and the same goes for randomness. What the sample size and the selection rule decide is not whether you may work quantitatively but which claims your results can carry. A census of one school supports description of that school and relational analysis within it; it does not support generalisation to other schools.
Stating that boundary explicitly is a strength in a proposal rather than a confession.
What is the difference between validity and reliability?
Reliability asks whether a measure gives consistent results each time it is used and whether scores are free of random error. Validity asks whether it measures the construct of interest at all, in four senses covering content coverage, agreement with an existing validated test, prediction of future criterion scores, and adequacy for the underlying construct.
A measure can be highly reliable and measure the wrong thing, which is why stability across time is evidence about the first and almost none about the second.
Can I report a mean for any variable I have coded numerically?
No, and this is one of the fastest errors for a reader to spot. Coding school type as one, two and three makes the numbers labels rather than quantities, so their average describes nothing. Means and standard deviations require at least interval measurement. For nominal data report counts and percentages, and if a single summary is needed, the mode.
Assessment move
Take one published quantitative article in your field and rebuild its measures section as a table: construct, instrument source, number of items, sample item, response scale, score computation, reliability. Any cell you cannot fill is a cell the authors left empty, and that is where you will learn what a complete measures section looks like.
Then build the same table for your own proposed study before you write a word of the methodology.
Working through Quantitative Research and Familial Influences in MEDD8001? Sia is AskSia’s AI Education tutor — ask any MEDD8001 Quantitative Research and Familial Influences question and get a clear, step-by-step explanation grounded in how MEDD8001 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
