QUIZ 1 Flashcards
how do we generate reliable knowledge?
by using the scientific method
observation–> ordering and classifying of facts–> generalizations –> hypothesis making –> testing –> verification –> knowledge
steps to designing research
problem–> question of interest –> specific predictions –> methods and research design –> data collection –> data analysis –> interpretation
research design example
migration and changes on agricultural patterns in Oaxaca, mexico
problem in Oaxaca
Is the arrival of remittances from migrants
changing the agricultural strategies of Zapotec
communities?
Oaxaca specific qs
what kind of changes are being implemented?
Oaxaca context
- place: mountains of Oaxaca
- Socioeconomic context: demographic
and economic collapse - Ecological issues: landscape ecology
oaxaca problem
Socioeconomic context results on
changes pressures over environment
Oaxaca variables
Relevant fields of inquire: agriculture
strategies, population, cash, commodities,
land cover
Oaxaca methods of data collection
1) Aerial picture analysis for land cover change
2) Demographic descriptive statistics and life stories
3) Tax records and mapping
4) Household income analysis
Question of Interest
oaxaca research design
Unit of analysis: individual/ household/
extended family
Timing of the process (1960s onwards), of the
research (seasonality?)
Scale: small community + multilocal
Sample: number of households
Question of Interest
Oaxaca results
- There is a clear process of forest
transition
People left
- Remittances are a fundamental part
of the local household economies
- Cultivar portfolio has changed (less
types of crops, less area devoted to
cultivation)
what is interdisciplinary research ?
Crosses traditional boundaries
between fields
research questions define…
context, scale, timing and history (process)
-variables, sample strategy, methods
classic research question problems
- Concept definition
- Required spatial scale of analysis
- Temporality (of event and of
research) - Goal definition
independent variable
initial variable of
which we know its changes
dependent variable
results on another
variable depending on the changes of the independent
variable
constants
value that doe snot change
either a reality or an assumption
process (diachronic studies)
the idea that things change across time
- time is an accumulation of points
process: consequences of synchronic studies
limits analysis of flows ( trends; predictions; patterns)
- idea of variability cannot be detached from the concept of process
process: questions and time
- temporality (diachronic/ synchronic)
- longitudinal vs cross-sectional
- repetitive relevance (ex: annual, seasonal)
- temporal scale (short to long term)
questions and space = scale of question
- Macro (relative to the question and
context) - Micro (relative to the question and context)
- Networked research (links between relevant
nodes) - Multiscalar research (links between different
scale levels)
breaking down research
variables (dependent/
independent)
Constants
Context
Process in time (history/ change)
Process in space
Evidence (data)
Sampling
what does time and space refer to in research Q?
- Demography across time
- Demography across space and
time - Redefining scale
what is replicability ?
The notion that same methods, same
locale, should generate the same results
how do you validate an interpretation ?
Validity would, then, depend on the
accumulation of such identical results
(statistical approach to the validation)
what is evidence?
data we produce
data we process
data we interpret
types of data
- quantitative / qualitative
- ‘objective’/ subjective
- artifacts (archaeological),
texts (interviews, novels, direct
observation), measurements..
question while sampling
- To who?
- To how many? –idea of
sampling- - How?
- What are the consequences of
each choice?
what is sampling theory?
The selection of some
part of the whole in
such a way that we
can use the part to
inform us about the
whole
what is probability sampling?
each element of population has equal chance of selection
define population
group of people, items or units under investigation
define census
information obtained by collecting
information about each member of a “population”
define sample
Obtained by collecting information only
about some members of a “population”
why do we use samples?
- Cost & time, or a census downright
impossible - Sampling provides adequate
information - Some tests are destructive (car
safety collision tests)
components of sampling
- Design (randomness, hierarchical,
snowball) - Size (representativity)
- Location (spatiality of the sampling)
- Composition (social variables):
gender, occupation, age, kin, status,
… - Awareness
how do you identify a ‘representative’ sample?
Sampling Theory (random)
- Each sample point must be independent
- Each sample point must have an equal
and independent probability of being
picked
- Adequate number of sample points
when to use random sampling
Natural Sciences prefer ‘Random’ or
‘Probability’ Sampling (otherwise results
may be biased, i.e., not representative of
population
why use non-random sampling?
Sometimes only biased samples are
available. Social sciences are conducive
to non-probability sampling: snowball
sampling, purposive, convenience
what are consequences of sampling?
- From the privileged sole informant,
to talking to everybody (from
minimal sample versus universe) - Reflecting about representativity
- Randomness versus purpose
types of random sampling
simple
systematic
stratified
cluster
simple random sample (equal chance)
Obtain a complete sampling frame
- Give each case a unique number
starting with one
- Decide on the required sample size
- Select that many numbers from a
table of random numbers
- Select the cases which correspond to
the randomly chosen numbers
systematic sampling (arranged in some order, first random, followed by k th)
Sample fraction
- divide the population size by the
desired sample size
- Select from the sampling frame
according to the sample fraction
- e.g sample faction = 1/5 means that
we select one person for every five in
the population
- Must decide where to start (start is
random)
stratified sampling
Premise - if a sample is to be
representative then proportions for various
groups in the sample should be the same
as in the population
Stratifying variable
characteristic on which we want to
ensure correct representation in the
sample
Order sampling frame into groups
Use simple random or systematic sampling
to select appropriate proportion of people
from each strata
cluster sampling
Involves drawing several different samples
by dividing a large geographic area into
smaller units
e.g., divide Montreal into boroughs
Select simple random samples from the
boroughs
start with large areas then progressively
sample smaller areas within the larger
types of non random sampling
snow ball
convenience
snow ball sampling
Identify possible informants by
asking our current informants
about suitable new subjects
Identification of networks
Ideal for specialized
communities
what kinds of of qs needs snowball sampling?
Questions on minorities or invisible
communities
- Questions on dispersed groups of
individuals (diaspora communities,
networks of specialized individuals,…)
- Questions on secretive of mistrustful
groups)
what is convenience sampling?
Glorified “do whatever you can”
what are control cases?
Chose two similar samples
- Proceed to the experiment with
one of them, leave the other as
an example of the initial
situation
why use control cases?
Asses change by, simultaneously
assessing lack of change
- understand the
mechanisms of change by
assessing two different processes
on identical locales
ethics: relevance
-Understanding the values of the
research site
- Understanding the
consequences of your research
- Conducting proper research
- Legal process
responsibilities in research
- To studied people and animals
(to subjects and context) - To scholarship and science
- To the public
research ethics
Research often confronts different
stakeholders interests
* Ethics as a complex field of
competing interests
* The researcher does not remain
outside of the game (becomes a
player or turned into one via
expectations)
how is data generated? (primary extraction)
- Observing social or biological behavior
- Interviewing
- Measuring frequencies
- Collecting samples
secondary treatment of data : processing
Statistics
- Discursive analysis
- Modeling
- Geographic Information Systems
different types of research methods
archival research or recollection of social data on the field
field data collection is gathered by either
surveys/ interviews or observation
composition of surveys and interviews
By structure
- structured, semistructured, unstructured
By theme
▪ Life stories, genealogies
▪ Free listing, triads pile sorting
▪ Diet breadth, income analysis
field observation
time allocation and participant observation (method and framework)
different fields of inquiry
- Demography
- Domestic Economy
- Ethnohistory
- Ethnobiology
- Health
what does interviewing consist of ?
Talking to people
- Opinion versus facts (interpretation)
- Narratives or points
- Practicalities: time, setting, themes
- Memory
types of questions
- Closed versus Open-ended Questions
- Closed questions includeYes/No responses, Likert
Scale questions, and Categorical Choices.
advantages of open-ended questions
When not all categories
are known
- Can answer in detail with
clarification
- Used if too many
categories
- Used if issue complex,
exploratory, preliminary
- Allows expressiveness
disadvantages of open-ended questions?
Worthless, irrelevant
responses possible
- Statistical Analysis
difficult
- Requires time to
respond
- Looks longer to
respondent
advantages of closed questions
Standardized
- Easier to respond to
- Easier to code
- Clearer about
meaning of question
- Better with sensitive
topics (multiple
choice)
disadvantages of closed questions
- Easy for respondent
to “just guess” - Respondent may not
find the right
category
wording to avoid in questions
avoid double-barrelled (and/or) and leading questions
order of interview questions
general –> specific –> open-ended and sensitive questions
historical data collections via…
written history (documents) = lit rev.
oral history
- interviews
-life stories
-genealogies
why is archival research important?
contextualization!! and historical data or state
how do you replicate archival research ?
citation of sources
contrasting sources
justification with data and source
what is a life story?
collection of recollections of personal historical narratives associated to individual past experiences
what is narrative analysis?
Local definitions of the key concepts (avoid
assumptions )
* Certain level of interpretation
* Narrative style and structure, presence of
metaphors
* Repetition across subjects
how do you replicate data from life story analysis?
Researcher interpretation
- Informants’ pollution
- Subjectivity
- Political motivations
- Competition
comparing life stories for internal consistency and reliability
- Compare versions of the same traditions told by
individuals from close groups - Compare accounts of stories affecting two groups
(migrations and wars) explained by individuals of
both groups - Discover regional trends by looking at whether
themes and events have entered into the oral
records of neighboring groups
social production of knowledge
- All landscapes are full of anthropogenic
features (resulting from human agency) - Social agency is informed by knowledge and
perceptions of reality - Knowledge is culturally organized
ethnobiological methods
- free lists
- triads
- pile sorting
- rankings
what does comparison of different social groups show?
unveil differences in how a specific cultural
domain is managed: occupation, gender, age, cultural or geographic origin
free list create?
spontaneous lists (of things, opinions…)
problems of free lists
- Over-differentiation and under-differentiation
(group versus species and subspecies) - Translation issues
-Previous knowledge of the question
-Expectations
triads
-attempt to identify classificatory logics
- provide three elements generated by the free lists and ask subjects to pair two of them (and explain why)
-need to pay attention to cultural and geographic context
classificatory rationality
morphological similarities
use
stories
ontological categories
pile sorting
-organize concepts in groups
-subdivide the groups (hierarchical clustering)
-interrogate about the logic behind distinctions
ranking
- Ask the informant to rank a data set
(provided by the researcher or produced by
the informant) according to a criteria
-Compare rankings depending on social
strata, cultural background, gender
difference between ranking and free lists
ranking has conscious classification (associated
to values or political views)
why do we observe behaviour ?
- Identification of behavioral patterns
- Understanding rationalities and constrains
behind those patterns - What people says is not always what people
do(ideal/bias/unconscious) - Memory is fickle
limits of interviewing
- history and memory
- self interested bias
- cultural expectations
- contradictory subjectivities
participant observation includes
long term field work, hang out/build trust, learning behavioural codes, describing everyday practices and learning local world view
contextual information includes
person, behaviour, setting (location), date & time, age, sex, household, and marital status
household econometrics
time allocation, income distribution, diet breadth,
the organization of time is significant, time sampling needs…
systematic following, self-administered, random spot sampling, people and places
types of sampling is according to
target !
focal sampling
single individual
several individuals, simultaneous behaviours
scan sampling
behaviour sampling
types of behaviour
sampling during a period
continuous sampling
instantaneous sampling
specific moments in succession
time can be a proxy for
productivity
efficiency
preference
problems of time
lying, overestimation, division of labor, cost of tools fabrication
key point of coding
defining categories
problems of coding
- simultaneity
- reliability
- context dependence
- mixing code categories
- classification of problems
what do we observed?
- Frequency (instances per unit time)
- Duration (length of single occurrence)
- Intensity (pace, useful for energetic
expenditure studies) - Sequence of behaviors (behavior flow) to
complete a task (steps in food preparation)
what is latency
the time between the end and start of a behaviour
goals of observation
- Sequence, duration, and frequency of
behavior - Understand the context of such behaviors
- Activities/ social indicators
- Identification of unconscious patterns and
trends (individual or collective)
common criticisms of observational methods
reductionism
focus on single issue
classification of behaviours is complicated
definition of categories, representativity of data collected, reactivity (researcher’s impact), size of observation + sample issues (randomized and size)
income distribution via interview
analysis of the composition of the income available to a household
source of the income
-type of activity, type of currency provides…
data on productivity, environmental impacts, trade and labour networks
-households connection to larger economic networks and inequalities
criticisms of income distribution
cash does not summarize wealth circulation, no data on redistribution, sensitive material, subject to high levels of occultation
analysis of socio-economic change
tradition: production and consumption
time devoted to production, actual profitability, changes on labour allocation viability
dietary breadth
collection of data on food consumption (who, what, how much, frequency)
diet breadth includes
diet composition
eating units
sharing networks
child-rearing units
dietary breadth correlates to socioeconomic variables
- Provides information on: nutrition, health,
production, trade - Differences between groups may point out
at inequalities or cultural preferences - Environmental consequences (diet emerges
from economic practices and these have
direct impacts on the environment)
demographics are made up of
structure and population dynamics
- size and territorial distribution of a population
- historical evolution of the population
demographic unit of reference
one population, analyzed via the family or the individual
types of demographic information
census : economic activity, level of eduction, ethnic group, civil status
Parrish register: marriage age, deaths
taxes: economic activities
situation of a population
Absolute size
▪ Abundance
▪ Size and settlement patterns
problems of population
useful surface, no total, problem of scale;
define the limits of the population, surface
(administrative units, property,… irrelevant with
groups that are not self-sufficient)
population structure
Age and sex (pyramid; in %). Information about history and
population
population pyramid trends
Broad base (young population, rapid growth),
▪ wider at the top (lack of generational renewal and population
reduction),
▪ similar values dif. age groups (stagnation).
▪ Strong differences indicate relevant past episodes
discursive analysis
- To detect patterns in usage and meaning
- To understand motivation and purpose
- To analyze internal structure and its
consequences
textual analysis or quantitative analysis includes
Content analysis Semantic networks Grounded theory
literary analysis
qualitative deconstruction
common issues of discourse analysis
selecting and contextualizing texts, coding and interpretation
literary analysis
Discourses and ideas as products of cultural
and historical context
¡ “Truth”: historical variable resulting from
uneven social relationships
¡ Common themes and rules and structures to
organize them
textual analysis: content analysis –> develop codebook
- Presence and prevalence of key words
- Relationships between texts, respondents, or
words (patterns) - Repetitions and connections between codes to identify the patterns
textual analysis: semantic networks
labeling to show relationships
(correspondence analysis =connections, hierarchical clustering =dependences)
study of the connections between nodes (concepts) more than the concepts themselves
textual analysis : grounded theory
–> Relationships amongst categories
–> Systematic coding of data
Topic of interest: describe lived experiences
- Diverse perspectives
- Multiple comparisons between data collected
- Unveil local meanings and local perspectives
- Focus often on a core category
example of combining methodologies
What do they say? (interviews)
What do they do? (behavioral
observation)
What do they eat? (diet breadth)
How do they pay for it? (income
analysis)
Social methodologies
-archival and bibliographical research
-interviews and questionnaires
-behavioral observation
analytical tools
population
the pool of individuals from which a statistical sample is drawn for a study.
census
procedure of systematically calculating, acquiring and recording information about the members of a given population
sample
a smaller, manageable version of a larger group. It is a subset containing the characteristics of a larger population
sampling frame
the actual set of units from which a sample has been drawn
random sample
in which each sample has an equal probability of being chosen
representative sample
is a sample from a larger group that accurately represents the characteristics of a larger population
data generating process
measurements taken from the real world (just a small glimpse) –> data
inferential statistics
allows you to draw conclusions based on extrapolations, and is in that way fundamentally different from descriptive statistics that merely summarize the data that has actually been measured.
examples of inferential statistical questions which amount to measuring differences
- What is the average level of life satisfaction in different
Canadian provinces? - Are successive (younger) cohorts in Quebec choosing to own
fewer cars?
causality
is a relationship between two events, or variables, in which one event or process causes an effect on the other event or process.
causality example
there is a positive correlation between ice cream sales and sunburns. Meaning, as ice cream sales increase, so do instances of sunburns.
causal salad
including confound while lacking a real causal model
inconvenient truths
-Covariates create confounds
-Prediction is not causal inference
-Data not enough
-Reproducibility not enough
sources of variation in causal diagrams
spatial, temporal, demographic variables
in an experiment; causal variable X is…
manipulated directly
a confounding variable causally affects both
X & Y
subjectivity of causality
-all conditions are causes
- often the difference between the “fundamental” one and
others is merely rhetoric or, rather, policy interest
measurement validity
how well your metric captures the
underlying concept you are trying to measure
internal validity
the degree to which the design of an
experiment controls extraneous variable, demonstrate cause-and-effect relationships
external validity
is the degree to which effects found in an
experiment generalise to other individuals, contexts, and outcomes.
For sampled studies, this means to times and places outside the
sampling frame ( can lack generalizability)
threats to external validity
1- interaction of selection and treatment: unrepresentative
responsiveness of the treated population
2 - interaction of setting and treatment: effect of the treatment
may differ across geographic or institutional settings
3 - interaction of history and treatment: effect of the treatment
may differ across time periods.
4- The effect may not persist, as individuals and institutions
adapt over time to the treatment.
5 - The treatment may be a “partial-equilibrium” effect (other
components of the sytem also undergo related changes,
reducing or eliminating the effect
considerations of experimental design
What is your treatment?
Who or what is the treatment group?
Who or what is the control group? How similar are they to
your treatment group?
How will you measure the treatment effect?
a classic experimental design ( pre-test/ post-test control group)
1 Random assignment to treatment and control groups
2 Control of the timing of the independent (treatment) variable.
3 Controls all other conditions under which the experiment
takes place.
4 Evaluate the differences-in-differences
what is an experiment?
a set of actions and observations, performed in the
context of solving a particular problem or question, to support or
falsify a hypothesis or research concerning phenomena
what are natural experiments
serendipitous situations in which
assignment to a treatment (or multiple treatments) and a control
group happens randomly and visibly, and outcomes are analysed
for the purposes of putting a hypothesis to a severe test
instrumented variation
if you are unable to experimentally vary the relevant variables,
researchers seek to find some variation in them that is driven by
factors that are clearly identified and understood. You can do this
through the use of an “instrument”
instrumented variation used when
-there is no fortuitous assignment into treatment/control
groups,
-there is no single natural driver of variation,
-and, in fact, there are confounding variables or two-way
causality that make causal identification difficult
natural vs instrumented natural experiments
Are subjects sorted unambiguously into different (discrete)
categories / treatments? (→ Natural experiment)
Or is the treatment composed of multiple influences, only one
of which (the instrument) is “random”, ie exogenous
to claim causality
time order - the cause must have occurred before the effects
co-variation (statistical correlation) - changes in the value
statistical control
we “control for” some variable or factor Z through statistical
adjustment, it means we try to take out the effect of Z on Y in
order to see what remains (which we may assume is due to X )
-This kind of “control” is done after the fact, during the statistics
phase, ie when the experiment or observation is done and the data
are in. = statistical adjustment
four fundamental confounds (directed acrylic graphs, DAGs)
the fork
the pipe
the collider
the descendent
covariate
an independent variable that can influence the outcome of a given statistical trial, but which is not of direct interest.
GIS software
ArcGIS
QGIS
GEOMEDIA- local gvts
Smallworld (GE)-used by utility companies
spatial features can be
discrete or continuous
discrete spatial features
houses, roads, wells = vector
continuous spatial features
rainfall, elevation = raster
discrete geographic features are better represented by
georelational vector data model (points, lines and polygons)
vector data rules
- thematic object forms its own layer ( roads separated from railways)
each layer can have only one type of feature ( can’t mix points with polygons)
continuous geographic features are better represented by
raster data model or grids & cells
an example of raster data model : the digital elevation model (DEM)
a digital terrain representation technique, where elevation values/topography are stored in raster cells
- useful for hydrological modeling
forms of raster data models
- aerial photographs (digital orthophoto quadrangle)
- satellite images
remote sensing and GIS
a form of ‘primary’ data collection
- can be used to collect information about objects on the ground using satellite or plane based sensors
pixel values in a raster image
valued between 0-255
0 - black
255- bright cell
colors are a proxy for the number values because
different land covers reflect different colors
why vegetation reflects near infrared
absorb red, green, and blue, to convert into food
infrared is all that’s left
spatial analysis in GIS
map projections
attribute data
cartography: making a map & choropleth maps
vector analysis with GIS using
buffers
overlays
-union
-clip
-intersection
buffer
polygon created by reclassification at a specified distance from point, line, area
overlay
places one ‘theme’ (e.g. soils) over another e.g (parking lots) e.g. check for soils which will cause problems of drainage for proposed parking lot
GIS analysis: buffer and overlay
buffers can be combined with polygon overlays in order to analyze spatial information
e.g. find all habitat areas of owls that are within 500 m of country roads
coordinate systems
(x,y) coordinate systems for drawn through the centre of the projection create new reference (x,y) for places in the globe
basic elements of a map
title, map features, legend, north arrow, scale bar, neat line
raster analysis
map algebra
- zonal
-focal
-local
-incremental
DEM-specific
-slope
-aspect
-cross-section
-inter-visibility
-hydrology