Geography — Research Methods Notes
Understanding Data in Research
Classification, Characteristics, Strengths and Limitations
Scope of this note
This note covers two foundational classifications used across geography, environmental science, and social-
science research: (1) Qualitative and Quantitative data, distinguished by the nature of information captured;
and (2) Nominal and Ordinal data, two sub-types of qualitative (categorical) data distinguished by the presence
or absence of a meaningful order.
Page 1 of 5
Geography — Research Methods Notes
1. Qualitative and Quantitative Data
All empirical research, whether in physical geography, human geography, or the social sciences more broadly,
rests on data — the raw facts, observations, and measurements collected about the phenomenon under study.
Before any statistical or interpretive technique can be applied, a researcher must understand what kind of data
has been collected, because the nature of the data determines which analytical tools are valid. The most
fundamental distinction in this regard is between qualitative and quantitative data.
1.1 Defining Qualitative Data
Qualitative data refers to information that describes qualities, characteristics, or attributes of a phenomenon
rather than its numerical magnitude. It is descriptive in nature and is typically expressed in words, categories,
images, or narratives rather than numbers. Qualitative data answers questions of 'what kind', 'which type', or
'how does it feel', rather than 'how much' or 'how many'.
In geography, qualitative data might include the type of soil in a region (laterite, alluvial, black cotton), the
perceived sense of safety among residents of a flood-prone settlement, the cultural significance attached to a
river by a community, or the classification of land use (residential, agricultural, forested, industrial). Such data
is often gathered through interviews, focus group discussions, participant observation, open-ended
questionnaires, oral histories, and textual or visual analysis of documents, maps, and photographs.
1.2 Defining Quantitative Data
Quantitative data, by contrast, refers to information that can be counted or measured and expressed
numerically. It answers questions such as 'how much', 'how many', or 'to what extent'. Quantitative data lends
itself to statistical treatment — it can be summarised through measures of central tendency (mean, median,
mode), dispersion (range, standard deviation, variance), and can be tested for relationships and significance
using inferential statistics.
Examples relevant to geography include rainfall recorded in millimetres, population density per square
kilometre, the number of households displaced by a cyclone, river discharge in cubic metres per second, or the
percentage of forest cover lost over a decade. Quantitative data is typically collected through structured
surveys, sensor and satellite readings, census records, government statistical databases (such as EM-DAT or
IDMC datasets used in disaster and displacement research), and controlled measurement instruments.
1.3 Key Points of Difference
Basis Qualitative Data Quantitative Data
Nature Descriptive, conceptual, subjective Numerical, objective, measurable
Expressed as Words, categories, themes, Numbers, counts, magnitudes
narratives
Central question “What kind / why / how” “How much / how many”
Collection tools Interviews, observation, open-ended Structured surveys, instruments,
surveys sensors, census
Analysis Thematic, content, narrative, Statistical: mean, correlation,
discourse analysis regression, GIS overlay
Sample size Usually smaller, in-depth Usually larger, for generalisability
Page 2 of 5
Geography — Research Methods Notes
Basis Qualitative Data Quantitative Data
Output Rich contextual understanding Generalisable, comparable figures
and trends
1.4 Advantages and Disadvantages of Qualitative Data
Advantages Disadvantages
Captures depth, context, and nuance that numbers cannot Difficult to generalise findings to a larger population
convey
Useful for exploring perceptions, meanings, and lived Analysis is time-consuming and can be labour-intensive
experience (e.g., a community's relationship with a river
or forest)
Flexible — allows new, unanticipated themes to emerge Prone to researcher subjectivity and interpretive bias
during research
Well suited to under-researched or culturally specific Harder to replicate exactly, reducing reliability across
contexts studies
Strong for generating hypotheses and for triangulating Not amenable to precise statistical testing or comparison
quantitative findings
1.5 Advantages and Disadvantages of Quantitative Data
Advantages Disadvantages
Objective, precise, and easily comparable across time Can strip away context, meaning, and lived experience
and space
Enables statistical testing, modelling, and prediction Risk of oversimplifying complex social or cultural
phenomena
Findings are more easily generalised to larger Quality entirely dependent on validity of measurement
populations instruments
Highly replicable — supports verification by other May force complex realities into rigid, predetermined
researchers categories
Efficient for handling very large datasets (e.g., census, Numbers can appear more 'objective' than they truly are,
satellite, climate data) masking bias in data collection
Applied note
In practice, strong research — including the kind used in Loss and Damage assessments or Forest Rights Act
spatial studies — often uses a mixed-methods approach: quantitative data (EM-DAT losses, GIS overlays,
census figures) establishes scale and pattern, while qualitative data (interviews, oral testimony, ethnographic
accounts) explains the human and cultural meaning behind those patterns.
Page 3 of 5
Geography — Research Methods Notes
2. Nominal and Ordinal Data
Nominal and ordinal data are two sub-categories of qualitative (categorical) data. Both classify observations
into discrete groups rather than measuring them on a continuous numerical scale, but they differ in one crucial
respect: whether the categories carry any inherent order or ranking. Together with interval and ratio data,
nominal and ordinal data form part of the four-level 'scales of measurement' framework first proposed by
psychologist S. S. Stevens (1946), which remains foundational to statistics and research methodology.
2.1 Nominal Data
Nominal data (from the Latin nomen, meaning 'name') consists of categories that have no logical order or
ranking — the categories simply label or name different groups. Numbers may occasionally be assigned to
nominal categories for coding purposes (e.g., 1 = male, 2 = female, 3 = other), but these numbers carry no
mathematical meaning; they cannot be added, subtracted, or ranked.
Examples of nominal data in geography and allied fields include: type of land use (agricultural, residential,
industrial, forest); religion or ethnicity of a surveyed population; type of soil (alluvial, laterite, black, red);
names of rivers, districts, or states; blood group; and gender. The only mathematical operations meaningfully
possible on nominal data are counting frequencies and determining the mode (the most frequently occurring
category).
2.2 Ordinal Data
Ordinal data also consists of categories, but here the categories can be logically ranked or ordered along some
dimension, even though the exact numerical distance between categories is unknown or not equal. The word
'ordinal' derives from 'order'. While we can say one category is 'higher' or 'lower' than another, we cannot say
precisely by how much.
Examples relevant to geography and disaster/policy research include: cyclone intensity categories (Category 1
to Category 5, as used for storms such as Phailin and Fani); levels of vulnerability to displacement (low,
medium, high, severe); Likert-scale survey responses (strongly disagree to strongly agree) on perceptions of
climate risk; socio-economic class (lower, middle, upper); and level of educational attainment (primary,
secondary, higher secondary, graduate, postgraduate). Ordinal data permits calculation of the median and
mode, and non-parametric statistical tests (such as the Mann-Whitney U test or Spearman's rank correlation),
but not the arithmetic mean in a strictly valid sense, since the intervals between ranks are not proven to be
equal.
2.3 Key Points of Difference
Basis Nominal Data Ordinal Data
Meaning of term From 'name' — categories are From 'order' — categories can be
simply labels ranked
Order among categories No logical or inherent order Clear, logical order or ranking exists
Distance between categories Not applicable Unknown / unequal, even if ranked
Examples Soil type, religion, river name, Cyclone category, vulnerability
gender level, Likert scale
Permissible statistics Frequency count, mode, chi-square Median, mode, rank-order
test correlation, non-parametric tests
Page 4 of 5
Geography — Research Methods Notes
Basis Nominal Data Ordinal Data
Mathematical operations Counting and classification only Ranking and ordering, but not
precise arithmetic
2.4 Advantages and Disadvantages of Nominal Data
Advantages Disadvantages
Simple, quick, and easy to collect and record Conveys no information about magnitude, rank, or
intensity
Ideal for classification and frequency-based summaries Limited statistical analysis possible (mode and
(e.g., land-use mapping) frequency only)
Easy for respondents to answer, improving response Cannot be used to measure change or degree over time
rates
Useful base layer for GIS categorisation (e.g., attribute Risk of oversimplifying complex identities into fixed
tables) labels
2.5 Advantages and Disadvantages of Ordinal Data
Advantages Disadvantages
Captures relative order or intensity (e.g., low to severe Exact difference between ranks is unknown or unequal
vulnerability)
More informative than nominal data while remaining Cannot calculate a true arithmetic mean
simple to collect
Well suited to perception and attitude-based research Interpretation of 'distance' between ranks can be
(Likert scales) subjective
Supports meaningful non-parametric statistical Fewer statistical techniques available compared to
comparison between groups interval/ratio data
Applied note
In cyclone and disaster-vulnerability research on Odisha — covering events such as Phailin, Fani, and Yaas —
nominal data (district name, land-use type) provides the classificatory base, while ordinal data (cyclone
category, vulnerability ranking) allows relative comparison of risk and impact across space, even though the
underlying interval-level wind-speed or damage data may also be collected separately for more precise
statistical work.
Page 5 of 5