|
| 1 | +--- |
| 2 | +title: "Output from sample grid" |
| 3 | +output: html_fragment |
| 4 | +--- |
| 5 | + |
| 6 | +```{r setup, include=FALSE} |
| 7 | +knitr::opts_chunk$set(echo = FALSE) |
| 8 | +``` |
| 9 | + |
| 10 | +# Sample output |
| 11 | + |
| 12 | +The results of the IC analysis are not displayed interactively but are bundled in an MS Excel file |
| 13 | +that can be downloaded (see tab `Analysis`). The results of all intermediate IC steps |
| 14 | +as described in [Burr, King, and Heckmann (2020)](https://doi.org/10.1080/14780887.2020.1794088) are |
| 15 | +contained. The main purpose of the software is to automate the cluster identification step of the IC |
| 16 | +procedure (see tab `Method`), which is a cumbersome and error-prone task if performed manually. |
| 17 | +Below, an extract of the analysis results for Sylvia's grid are shown. For a more detailed epxlanation, |
| 18 | +however, please refer to our publication. |
| 19 | + |
| 20 | +## Input data and results |
| 21 | + |
| 22 | +The input file for Sylvias's grid is available for download on the `Analysis` tab. Figure 1 shows the |
| 23 | +raw grid data. |
| 24 | + |
| 25 | +{width="60%"} |
| 26 | + |
| 27 | +Figure 2 displayes the network graph of related contructs and discovered construct cliques, the most |
| 28 | +relevant part of the output, which is subsequently used for interpretation in the next section. |
| 29 | + |
| 30 | +{width="60%"} |
| 31 | + |
| 32 | +## Interpretation |
| 33 | + |
| 34 | +Psychologically relevant information can be obtained from the interpretation of the network graph in |
| 35 | +Figure 2. What follows is a shortened example. More comprehensive examples are outlined in our |
| 36 | +[publication](https://doi.org/10.1080/14780887.2020.1794088). |
| 37 | + |
| 38 | +In the resulting diagram for Sylvia's grid in Figure 2, a construct is indicated by a circle, with |
| 39 | +(+) denoting the preferred and (-) the non-preferred pole. The diagram shows three clusters (also |
| 40 | +called cliques), indicated by the colored hulls around several constructs. In Sylvia's case, the |
| 41 | +three clusters are highly overlapping. Two of these are of particular interest, sharing a 'core' of |
| 42 | +three constructs -- '(+) Wild, free *vs* (-) controlled, contained', '(+) Massive sense of space, |
| 43 | +expansive *vs* (-) railed-off, small world' and '(+) Freedom, wildness *vs* (-) conventional', with |
| 44 | +(+) indicating the preferred and (-) the non-preferred pole. In one cluster, these three constructs |
| 45 | +are strongly associated with '(+) Verdant *vs* (-) dead, nothing thriving'; the association between |
| 46 | +her preferred poles suggests that she is drawn to places that are thriving and green, wild and |
| 47 | +expansive, as opposed to those which lack life, are small-scale, controlled and conventional. |
| 48 | +However, these three constructs share another cluster with the construct '(+) Cosy *vs* (-) |
| 49 | +Chaotic', where 'cosy' is her preferred pole. In this cluster, however, her desires for the wild, |
| 50 | +free and expansive appear to be in tension with her desire for the 'cosy', as they are aligned with |
| 51 | +her non-preferred pole 'chaotic'. The attraction of wild, free spaces for Sylvia is therefore not |
| 52 | +straightforward. |
| 53 | + |
| 54 | +The third cluster includes the '(+) Verdant *vs* (-) dead, nothing thriving' construct, which is |
| 55 | +here associated with '(+) Exciting, a lot going on *vs* (-) flat, unvarying, depressed, |
| 56 | +unenergetic', '(+) Dramatic *vs* (-) unvarying, goes on and on' and '(+) Variable *vs* (-) doesn't |
| 57 | +change'. This suggests that to Sylvia 'verdant' spaces are also full of excitement, drama and |
| 58 | +variability- they are full of life in these ways. However, the fact that these three constructs do |
| 59 | +not cluster with the wild/expansive/freedom constructs indicates that they constitute a somewhat |
| 60 | +separate idea for her. A 'wild' space for her need not be 'exciting', for example, although a |
| 61 | +'verdant' space is likely to be both exciting and wild. Interpretive clustering therefore gives us |
| 62 | +insight into some of the complexity of Sylvia's construing. |
| 63 | + |
| 64 | +```{r include=FALSE} |
| 65 | +library(openxlsx) |
| 66 | +library(OpenRepGrid.ic) |
| 67 | +
|
| 68 | +file <- system.file("extdata", "sylvia.xlsx", package = "OpenRepGrid.ic") |
| 69 | +x <- read.xlsx(file) |
| 70 | +s <- calculate_similarity(x, min_matches = 6) |
| 71 | +``` |
| 72 | + |
| 73 | +## Underlying calculations |
| 74 | + |
| 75 | +Also contained in the Excel ouput are the matrices generated in each step of the IC method. Some |
| 76 | +details are explained on the `Method` tab. As an example, Table 1 shows the number of matches between constructs, with constructs reversal |
| 77 | +beeing allowed. Thus, Construct 2 matches on three elements with Construct 1, on five elements with |
| 78 | +Construct 3, on three elements with Construct 4, and so on. |
| 79 | + |
| 80 | +*Table 1:* Number of matches between constructs. |
| 81 | + |
| 82 | +```{r} |
| 83 | +knitr::kable(s$R, table.attr = "style='width:60%;'") |
| 84 | +``` |
| 85 | + |
| 86 | +The matrix of constructs relations is inferred from the matrix of construct matches, here, using a |
| 87 | +minimum of 3 matches. The result is displayed in Table 2. A value of `1` denotes a positive, a value |
| 88 | +of `-1` a negative construct relation. This matrix is used as a basis to build the network graph from |
| 89 | +Figure 2. |
| 90 | + |
| 91 | +*Table 2:* Relations between constructs. |
| 92 | + |
| 93 | +```{r} |
| 94 | +knitr::kable(s$D) |
| 95 | +``` |
| 96 | +The examples above give a first insight into the generated output. Yet, to fully understand and work |
| 97 | +with the output, it is recommended to consult [Burr, King, and Heckmann (2020)](https://doi.org/10.1080/14780887.2020.1794088) |
| 98 | +which outlines the IC method and gives a more thorough interpretation example. |
0 commit comments