Chapter 8 Longitudinal Research

Longitudinal Research

Collection of data at different points in time

Inter-cohort Differences

Comparing different age cohorts at one point in time

Cross-sectional Research

Data collected from a representative sample at only one point in time

Longitudinal Research in Experiments

Delayed assessment is essential to measure the duration of impact of media effects

Trend Studies

Different people are asked the same question at different points in time

Period Effects

Influences associated with each particular time period

Cohort Effects

Influences associated with members in a certain birth cohort

Age Effects

Influences produced by growing older

Intra-cohort study

Reveal changes in a single cohort from one time to another

Follow-back panel

Selecting a cross-sectional sample in the present and using archival data to create the longitudinal dimension of the study

Catch-up Panel

Selecting a cross-sectional study done in the past and locating all possible units of analysis for observation in the present -lack enough identifying information to track down subjects -comparability of measures -limited to original variables measure

Cohort Analysis

The study of specific populations, usually all those born during a given period, as they change over time

Cohort

any group of individuals who are linked in some way or who have experienced the same significant life event within a given period

Cross-lagged Correlations

information about two variables is gathered from the same sample at two different times

Special Panel Designs

1. Retrospective panel 2. Follow-back panel 3. Catch-up panel

Causal Explanations

1. Ideographic 2. Nomothetic Different approaches have different priorities

Advantages and Disadvantages of Follow-back panel

+Changes that occurred over a long period of time can be analysed in a short period +Studying dwindling populations -Archives do not contain many variables that interest mass media researchers -May not represent all possible entities

Advantages and Disadvantages of Trend Studies

+Describe long-term changes in a population +Can be based on a comparison of survey data originally constructed for other purposes -Results are only as good as underlying data -Must be based on consistent measures

Advantages and Disadvantages of Cohort Analysis

+Highly flexible +Can be used with either original or secondary data -Specific effects of age, cohort and period are difficult to untangle through purely statistical analysis of a standard cohort table -Sample mortality

Longitudinal Research on the Internet

+Large number of individuals can be recruited +Find people from low-incidence groups -May not be representative -Sample mortality -Unable to verify identity of respondent

Advantages and Disadvantages of Combining Quantitative and Qualitative Data

+Qualitative helps in interpretation of quantitative +Suggest new hypotheses for future waves -Increase time and effort to analyse data

Advantages and Disadvantages of Panel Studies

+Useful to answer the dynamics of change +Repeated contacts reduce suspicions which may help yield more data than initial encounters +Allow for competing stimuli on the subject -Panel members are difficult to recruit -Sample mortality -Sensitisation Effect -Respondent error

Cohort Analysis Table Comparisons

1. Inter-cohort differences (down the column) 2. Trends at each age level (across a row) 3. Intra-cohort study (diagonal to the right)

Advanced Statistical Techniques

1. Path analysis 2. Log-linear models 3. LISREL (LInear Structural RELations)

Using Different Samples

-Bad science -Convert the data into z-scores to make it valid

Panel Studies

-Measuring the same sample at different points in time -Unlike trend studies, panel studies can reveal information about both net change and gross change in the dependent variable

Ideographic Explanation

-attempts to identify all influences -for specific entities (i.e. particular groups or individuals) and phenomena [particular]

Nomothetic Explanation

-identify a few causal factors -for many entities or phenomena (i.e. general trends) -probabilistic (uncertain and incomplete) -parsimonious (frugal, just enough)

Nomothetic Approach

-nomothetic: from ancient Greek nomothetikos, meaning "law" or "lawgiver" -predominant form of social science research -supported by law of diminishing returns

Cohort Analysis Table

-usually involves data from more than one cohort -interval between 2 periods of measurement corresponds to the age class intervals

Variation Effects

1. Age 2. Cohort 3. Period

Minimising Sample Mortality in Panel Studies

1. Cash incentive 2. Establish the credibility and value of the research project 3. Gather detailed information about the panel member (whereabouts, relationships) 4. Contact panel member between data collection waves 5. Give panel members a contact card to call if they change addresses

Types of Panel

1. Continuous 2. Interval Usage depends on purpose of the study

Conditions for Causality

1. Time order (cause precedes effect) 2. Association 3. Non-spurious (rule out alternative causes)

Types of Longitudinal Studies

1. Trend study 2. Cohort Analysis 3. Panel Study

Birth cohort

A group born within a year or time period

Interval Panel

Members agree to complete a certain number of measurement instruments only when the information is needed

Continuous Panel

Members who report regularly

Retrospective Panel Study

Respondent is asked to recall facts or attitudes from the past and these recalled factors are then compared with a later measure of the same variable, thus producing an instant longitudinal design Weigh the cost savings with possible losses in accuracy