The aim is to describe the development of a middle-range theory by using an inductive-deductive approach. A theory of well-being supporting physical environment of home-dwelling elderly is used as an example. The inductive-deductive theory development process is described through four different phases: (1) the creations of concepts were described inductively through concept synthesis, (2) relationships between the concepts were examined to set up a hypothetical model, (3) hypotheses were set up to verify the concepts and to test hypothetical models, and (4) the verification and presentation of the theory.
Nursing science theories can be divided into metatheories, conceptual models, and middle-range and small theories [
Nursing science theories can be developed using either an inductive or a deductive approach. When developing a theory, these stages often occur in turns, and the development process may begin using either an inductive or a deductive approach, depending on the nature of previous research data. An inductive approach to developing a theory is used when the aim is to form a theory on a subject on which little information is available, if the information is fragmentary or if the aim is to find a new viewpoint on the subject. When forming a theory using inductive methods, individual observations are used as the basis. Based on several individual observations, conclusions may be drawn and a theory constructed [
Theories can be classified based on their type or level [
Developing nursing science theory through research has become increasingly more common since the 1980s. However, studies on the subject were not published in scientific journals until the 1990s. Around that time, discussion arose on the significance of different-level theories in the advancement of science [
The purpose of this paper is to describe the development of a middle-range theory by using an inductive-deductive approach. The example theory has been developed according to the following phases: (1) selection of concepts and their synthesis, (2) definition of relationships between concepts, (3) setting and testing hypotheses, and (4) presentation and verification of the theory. The first two phases represent an inductive approach and the last two a deductive approach. The phases are presented in Figure
The phases of forming a theory, adapted from Lauri & Kyngäs 2005 [
The study used as an example looks at well-being from the viewpoint of self-perceived or subjective well-being, meaning that the well-being of the elderly is evaluated by the elderly themselves. Following the principle of inductive theory formation, well-being was not defined in advance: the definition was based on the experiences of the study subjects. The starting point was that the environment was considered as a source of well-being, with the elderly seen as fulfilling their needs and the environment as a resource that contributes to well-being.
The main concepts of the well-being supportive physical environment include an environment that enables safe activity, a pleasant physical environment, and Northern environment. An environment that enables safe activity comprises both safety at home and immediate surroundings that enable safe mobility. At home, important aspects include the safety of stairs and steps, reducing the need to reach or climb, the use of various support rails, and making floors and the bathroom less slippery. In the immediate surroundings of the home, ensuring safe mobility calls for well-maintained traffic routes and the use of technical aids. A pleasant physical environment consists of tidiness at home and in its immediate surroundings, closeness to natural environment, and opportunities for various activities. Natural environment areas, such as parks and gardens, are popular meeting places that provide an opportunity to interact with other people. They are also popular among the elderly for exercise and relaxation. A pleasant living environment also includes possibilities to engage in various activities. Several factors associated with Northern environment, such as climate and availability of services, are related to the well-being of the elderly. For example, various cold- or heat-related symptoms during the winter or summer make everyday coping more difficult and weaken the perceived state of health of the elderly [
Formulation of concepts was the first phase of theory development and concepts were formed through concept synthesis. The aim of this was to observe the concept in a systematic and disciplined manner by an inductive approach [
The concept synthesis is based on a qualitative approach, with interview data as a basis. The interview data were gathered from home-dwelling elderly people over the age of 65 in Northern Finland (
After transcribing the interview data, concept synthesis was used to construct concepts based on the information gathered from individual observations by grouping and organising information regarding the phenomenon [
Through concept synthesis, the aim was to construct models representing the phenomenon under study in a condensed form by producing concepts that describe the phenomenon. The analysis was begun by selecting a unit of analysis—a letter, a word, a sentence, or a longer semantic unit, depending on the aims of the study at hand [
In order to form concepts, synonymous statements were merged to form subconcepts, and these were given names that described them well [
An example of a hypothetical model.
In a deductive phase the hypotheses for testing the theory are set and tested [
In order for hypotheses to be tested, there must be an instrument that can be used to test the concepts presented as hypotheses [
The reliability of a theory calls for a valid instrument. Pretesting plays a significant part in assessing the validity of the indicator [
To test the theory, a pretested version of the instrument was sent in first phase to 500 home-dwelling elderly people between ages 65 and 74 in Northern Finland. The stratified random sampling method was used. Based on the data (
The principal component analysis was carried out in four phases. First, the correlation coefficients and covariances between items that were high enough to allow analysis (>0.30) were calculated. In the second phase, estimated factor loadings of the principal components/factors were obtained for the matrix formed based on the correlation coefficients and covariances. Items with loadings over 0.40 were included in the factors. In the literature, recommendations for factor loadings to be included in the analysis vary between 0.35 and 0.55 [
After the analysis, the goodness of the principal components was estimated in terms of content and based on the factor loadings of the items. The goodness of individual items loaded on the principal components was investigated using communalities. Communality reveals how strongly the item is loaded on the principal component, giving an item with strong loading a communality value close to 1. Items not fulfilling the requirements of goodness may be removed based on low communalities. The main principle is that the communality of an item should be >0.30. The communalities of all the items exceeded this value, indicating that they measure the factors in a relatively reliable manner. All the factors were relevant in terms of content, though some differed from the hypothetical model. For example, factors related to living comfort of the natural environment and tidy environment were loaded on the same factor. All the Northern environment factors that impair mobility were also loaded on the same factor. These alterations were made in the hypothetical model for the purposes of confirmatory factor analysis.
Confirmatory factor analysis (CFA) was used to ensure that the data supported the models constructed through principal component analysis [
After CFA, all factors, that is, principal components, were named. They can be named according to the item or items with the highest loading and according to the processes contributing to the origin of the factors. However, from a theory-developing point of view a good alternative is to base the naming on theory. In this study, naming was based on the results of the first phase of the study, during which theoretical concepts were formed. The theoretical concepts describing the physical environment as well as their alpha values are presented in Table
Names, alpha values, and number of items of the concepts describing physical environment
Concept | Cronbach’s alpha | Number of items |
---|---|---|
Alterations in the home facilitating domestic activities | .94 | 3 |
Natural environment contributing to living comfort | .82 | 3 |
Factors of Northern environment complicating mobility | .84 | 3 |
Temperature as a factor complicating life | .68 | 2 |
Natural environment as a pleasant location for exercise | .80 | 3 |
Darkness as a factor promoting depression | .93 | 2 |
Well-maintained traffic routes facilitating mobility | .67 | 2 |
Availability of services as a prerequisite for living at home | .65 | 2 |
Meeting familiar people in a natural environment | .60 | 2 |
Moving in the immediate vicinity | .50 | 2 |
A tidy environment contributing to living comfortably | .62 | 2 |
Developing a theory is a process that should not end with presenting the theory; the theory should be tested at different time points and with different target groups. Based on confirmatory factor analysis results the theory of the well-being supportive physical environment has been verified in three different data during years 2005–2012 (Table
The results of confirmatory factor analysis (data 1–data 3) as indices of relevance to the data.
Model and data | Indices of relevance | ||||||||
---|---|---|---|---|---|---|---|---|---|
|
df |
|
GFI | AGFI | RMR | NFI | CFI | RMSEA | |
Northern environment, four-factor model | |||||||||
Data 1: Northern Finland 75–85 | 49 | 21 | <.00 | .98 | .96 | .05 | .97 | .98 | .05 |
Data 2: Southern Finland 65–74 | 40 | 19 | <.00 | .98 | .95 | .06 | .97 | .98 | .05 |
Data 3: Northern Finland 65–74 | 340 | 21 | .01 | .98 | .95 | .05 | .97 | .99 | .05 |
An environment that enables safe activity | |||||||||
Data 1: Northern Finland 75–85 | 8 | 11 | .68 | .99 | .99 | .3 | .99 | 1 | <.00 |
Data 2: Southern Finland 65–74 | 10 | 11 | .57 | .99 | .98 | .02 | .99 | 1 | <.00 |
Data 3: Northern Finland 65–74 | 7 | 11 | .80 | .99 | .99 | .04 | .99 | 1 | <.00 |
A pleasant physical environment | |||||||||
Data 1: Northern Finland 75–85 | 60 | 28 | <.00 | .98 | .96 | .04 | .96 | .98 | .05 |
Data 2: Southern Finland 65–74 | 75 | 29 | <.00 | .96 | .93 | .06 | .94 | .96 | .07 |
Data 3: Northern Finland 65–74 | 52 | 30 | .008 | .97 | .95 | .02 | .95 | .98 | .05 |
A widely used criterion in evaluating a theory is applying the theory in practice [
The theory was tested with confirmatory factor analysis which showed that the validity and reliability of the theory was relatively good. The statistical values as well as the indices of relevance to the data presented as a result of the analysis are sufficient in all the models tested. Using factor analysis in testing the theory was found to be a successful solution, and the indices obtained in the analysis confirmed the hypothetical models formed by the researcher through qualitative analysis.
The theory was tested with confirmatory factor analysis which showed that the validity and reliability of the theory were relatively good. The statistical values as well as the indices of relevance to the data presented as a result of the analysis are sufficient in all the models tested. Using factor analysis in testing the theory was found to be a successful solution, and the indices obtained in the analysis confirmed the hypothetical models formed by the researcher through qualitative analysis. Cronbach alpha values are shown in Table
This paper also gives an example of how to develop and verify an inductive-deductive nursing science theory. Theory development is multiphases process. Here the theory is based on inductive approach which provides perspective of elderly people. After inductive phase there was a need to test and verify the theory by deductive approach. Advanced statistical methods enable us to study the relationships between concepts and build the theories. Middle-range nursing theories are needed to develop the practice and nursing science itself.