QUANTITATIVE METHODS IN MEDIA AND COMMUNICATIONS
LECTURE WEEK 1
Qualitative methods – interpretative, deep, nuances fluffy pretentious, feely
Quantitative methods – deductive, generalizable, evidence rigid, meaningless, numbers
5-ever
Theory – an interrelated set of variables formed into hypotheses that specify the
relationship among variables, with the purpose of explaining natural phenomenon
Theoretical rationale – provides an explanation or prediction about why and how variable X
would influence variable Y
Qualitative RQs Quantitative RQs
Aim To discover To test (hypotheses)
To seek to understand To examine relationship
To explore variables
To describe To compare
To describe
Type Open questions More narrow questions
Examples How… To what extent…
What is meaning… What effect…
…causes…
THE USE OF THEORY IN QUANTITATIVE RESEARCH
Variable – a characteristic or attribute that varies among the people that are being studies
Independent Variables (IV) – those that probably influence or affect outcomes (treatment,
manipulated, antecedent, predictor) causes, influences something else
Dependent Variables (DV) – those that are the presumed result of the influence of the IV
(criterion, outcome, effect) influenced by the independent variable
Hypotheses versus Opinions – main differences
- Hypotheses are tested with statistics, opinions are not
- Hypotheses can then be answered by saying TRUE or FALSE
- Hypotheses are followed up by data
TWO TYPES OF HYPOTHESES
1. Null hypothesis (H0)
- No difference or change
- Never stated, always implied
2. Alternative hypothesis (H1)
- Statement of prediction
- Actual research hypothesis
The goal of hypothesis testing is to reject one hypothesis and accept the other
TYPES OF H1 HYPOTHESES
- Directional hypothesis – difference or effect in particular direction (e.g. Media
violence increases mood)
, - Nondirectional hypothesis – difference or effect but not in particular direction, avoid
when possible (e.g. media violence influences mood)
VARIABLES THAT INFLUENCE THE RELATIONSHIP OF IV AND DV
Spurious relationship – a relationship in which the independent and dependent variable
seem related, but are in fact not
Confounding variables – variables that are not measured but might influence or explain the
observed relationship (e.g. the more chimneys, the more pregnant women the
confounding variable here is the number of citizens; the more citizens, the more houses (and
chimneys) and the more pregnant women within a community)
According to Cresswell, spurious and confounding relationships are the same this is NOT
true in reality, the independent variable and dependent variable in a spurious relationship
are completely unrelated, there is nothing that links them in a confounding relationship,
however, the independent variable and dependent variable are related through a third
variable that links them e.g. the number of citizens predicts both the number of chimneys
and pregnant women
Control variables – a special type of independent variable that researches measure because
they potentially influence the dependent variable (usually demographics such as age)
Intervening/mediating variables – stand between the independent and dependent
variables, and ‘mediate’ the effects of the independent variable on the dependent variable
tell you how an independent variable affects the dependent variable
Moderating variables – (e.g. gender) affect the relationship between the independent and
the dependent variable, such that the effect may be present for one groups (e.g. males) but
not another (e.g. females) moderation variables tell you for whom an independent
variable affects the dependent variable
LECTURE WEEK 2
Data are only good as the instrument that you use to collect them “good science” also
includes careful operationalization (e.g. defining, refining, selection of measurements)
INTRODUCING SCALES
- The variables (characteristics or attributes that varies among the people that are
being studied) in your research question(s) and hypotheses may be abstract concepts
(e.g. press freedom, materialism, life satisfaction, friendship quality, addiction)
- Abstract concepts cannot be measured with one question, some characteristics can
(e.g. gender, sex, educational level, ethnicity, nationality, age)
- We can ask and expect a response to questions like: what gender do you identify
with? What is the highest educational level you achieved? But it makes littles sense
to ask What is your level of materialism? Or How high is your friendship quality?
- To measure abstract concepts we generally use a set of questions (scale)
- Scales contain multiple items, being questions or statements, which participants
have to react to
- For each item of a scale, an identical number of answer options should be provided
- For the MVS-C there were 4 answer options: (1) no, not at all, (2) no, not really, (3)
yes, a little, and (4) yes, a lot
,SCALES SCORES CAN BE CALCULATED IN 2 WAYS
- The scores on the separate items (e.g. questions/statements) can be summed or
averaged depending on how the final variable will be used
- Sum scores are used, when the scale is meant to categorize the participants (e.g.
intelligence, autism, addiction, total number of hours spent with media)
- In general, researchers prefer to use average scale scores, because this allows us to
interpret the scale score with the original answer options
- Asking ourselves whether the participant is materialist (e.g. the answer could be 3
out of 4) the child would be a little materialistic
- Importantly, some scales combine positive and negative items in such cases, the
negative items need to be recoded before the scale score is calculated
- The eventual scale score is calculated by determining the average of the original
positive items and the recoded negative items e.g. when I look at the world, I
don’t see much to be grateful for & long amounts of time can go by before I feel
grateful to something or someone
3 CRITERIA FOR A GOOD OPERATIONALIZATION
- A reliable measure – the measure always measures the same
- A valid measure – you measure what you wanted to measure
- An objective measure – the measure measures the same when you conduct the
study or when your neighbor does it
always check the scales’ internal consistency and content validity – the other criteria,
unfortunately, can often not be verified
Researchers that introduce new scales will provide the following information:
- The number of items (questions/statements)
- The content of the items (questions asked)
- The number of answer options (how many answer options)
- The content of the answer options (what are the possible answers)
- Whether some items need to be reverse scored
- The reliability of the scale
- The validity of the scale
when designing the questionnaire for a new study, “good” researchers will include pre-
existing scale that have been proven reliable and valid
RELIABILITY OF SCALES
- Two frequently used indicators of a scale’s reliability are internal consistency and
test-retest reliability
- Internal consistency is the degree to which the items that make up the scale are all
measuring the same underlying attribute when talking about reliability, this is
usually what researchers refer to
- Cronbach’s alpha is a measure of internal consistency the statistic indicates the
average correlation among all of the items that make up a scale
- Cronbach’s alpha ranges between 0 and 1.00 values above 0.70 are considered
acceptable (reliable scale) values above 0.80 are preferable
Choosing between scales – Reliability
, The test-retest reliability (temporal stability) of a scale is assessed by
administering it to the same people at multiple occasions, and calculating the
correlation between the two scores obtained
- Small correlations (r = 0.10 to 0.29) indicate low test-retest reliability
- Medium correlations (r = 0.30 to 0.49) indicate moderate test-retest
reliability
- Large correlations (r = 0.50 to 0.99) indicate high test-retest reliability
- Zero or negative correlations are a sign of poor test-retest reliability
Validity of scales
- The validity of a measure refers to the degree to which it measures what it is
supposed to measure
- There are 3 types of validity: content validity, criterion validity, and construct
validity
Content validity refers to the extent to which the items of the scale are representative of
the entire domain the scale intent to measure do the questions capture the concept well?
Some scales consist of subscales, in such case the items should capture the designated
subconcept
Criterion validity concerns the relationship between scale scores and some specified,
measurable criterion – this is usually assessed through a scale’s correlation to other scales
intended to measure the same or a similar concept (e.g. new materialism scale is measured
against other known materialism scales)
- Small correlations (r = 0.10 to 0.29) indicate low validity
- Medium correlations (r = 0.30 to 0.49) indicate moderate validity
- Large correlations (r = 0.50 to 1.00) indicate high validity
Construct validity involves testing a scale not against a single criterion but in terms of
theoretically derived hypotheses concerning the nature of the underlying variable or
construct the scale should be related to other concepts it is known to be related to
(convergent validity), and unrelated to concepts it is known to be unrelated to (discriminant
validity) e.g. materialism should be positively correlated with being a boy, and negatively
with life satisfaction, both refer to convergent validity
Writing/improving survey questions
Advantages Disadvantages
Open-ended questions 1. Respondents can 1. Takes longer to
freely respond in answer questions
unique manner 2. Responses might not
2. Respondents might be legible
reveal unexpected
insights
Closed-ended questions 1. Respondents can 1. Respondents are
(used more in surveys) answer questions limited in their
quickly responses
2. Respondents call be 2. Respondents might
