R/automated_readability_index.R

In trinker/qdap2: Bridging the gap between qualitative data and quantitative analysis

#' Readability Measures
#' 
#' \code{automated_readability_index} - Apply Automated Readability Index to 
#' transcript(s) by zero or more grouping variable(s).
#' 
#' @param text.var The text variable.
#' @param grouping.var The grouping variables.  Default NULL generates one 
#' output for all text.  Also takes a single grouping variable or a list of 1 
#' or more grouping variables.
#' @param rm.incomplete logical.  If TRUE removes incomplete sentences from the 
#' analysis.
#' @param \ldots Other arguments passed to \code{\link[qdap]{end_inc}}.
#' @return Returns a dataframe with selected readability statistic by grouping 
#' variable(s).  The \code{frey} function returns a graphic representation of 
#' the readability as well as a list of two dataframe: 1) \code{SENTENCES_USED} 
#' and 2) \code{SENTENCE_AVERAGES}.
#' @rdname Readability
#' @section Warning: Many of the indices (e.g. Automated Readability Index) 
#' are derived from word difficulty (letters per word) and sentence difficulty 
#' (words per sentence).  If you have not run the sentSplit function on your 
#' data the results may not be accurate.
#' @references Coleman, M., & Liau, T. L. (1975). A computer readability formula 
#' designed for machine scoring. Journal of Applied Psychology, Vol. 60, 
#' pp. 283-284.
#' 
#' Flesch R. (1948). A new readability yardstick. Journal of Applied Psychology. 
#' Vol. 32(3), pp. 221-233. doi: 10.1037/h0057532.
#' 
#' Gunning, T. G. (2003). Building Literacy in the Content Areas. Boston: Allyn 
#' & Bacon.
#' 
#' McLaughlin, G. H. (1969). SMOG Grading: A New Readability Formula. 
#' Journal of Reading, Vol. 12(8), pp. 639-646. 
#' 
#' Senter, R. J., & Smith, E. A.. (1967) Automated readability index. 
#' Technical Report AMRLTR-66-220, University of Cincinnati, Cincinnati, Ohio.
#' @keywords readability, Automated Readability Index, Coleman Liau, SMOG, 
#' Flesch-Kincaid, Fry, Linsear Write
#' @export
#' @examples
#' AR1 <- with(rajSPLIT, automated_readability_index(dialogue, list(person, act)))
#' htruncdf(AR1,, 15)
#' AR2 <- with(rajSPLIT, automated_readability_index(dialogue, list(sex, fam.aff)))
#' htruncdf(AR2,, 15)
#' 
#' CL1 <- with(rajSPLIT, coleman_liau(dialogue, list(person, act)))
#' head(CL1)
#' CL2 <- with(rajSPLIT, coleman_liau(dialogue, list(sex, fam.aff)))
#' head(CL2)
#' 
#' SM1 <- with(rajSPLIT, SMOG(dialogue, list(person, act)))
#' head(SM1)
#' SM2 <- with(rajSPLIT, SMOG(dialogue, list(sex, fam.aff)))
#' head(SM2)
#' 
#' FL1 <- with(rajSPLIT, flesch_kincaid(dialogue, list(person, act)))
#' head(FL1)
#' FL2 <-  with(rajSPLIT, flesch_kincaid(dialogue, list(sex, fam.aff)))
#' head(FL2)
#' 
#' FR <- with(rajSPLIT, fry(dialogue, list(sex, fam.aff)))
#' htruncdf(FR$SENTENCES_USED)
#' head(FR$SENTENCE_AVERAGES)
#' 
#' LW1 <- with(rajSPLIT, linsear_write(dialogue, list(person, act)))
#' head(LW1)
#' LW2 <- with(rajSPLIT, linsear_write(dialogue, list(sex, fam.aff)))
#' head(LW2)
automated_readability_index <-
function(text.var, grouping.var = NULL, rm.incomplete = FALSE, ...) {
    if(is.null(grouping.var)) {
        G <- "all"
    } else {
        if (is.list(grouping.var)) {
            m <- unlist(as.character(substitute(grouping.var))[-1])
            m <- sapply(strsplit(m, "$", fixed=TRUE), function(x) {
                    x[length(x)]
                }
            )
            G <- paste(m, collapse="&")
        } else {
            G <- as.character(substitute(grouping.var))
            G <- G[length(G)]
        }
    }
    if(is.null(grouping.var)){
        grouping <- rep("all", length(text.var))
    } else {
        if (is.list(grouping.var) & length(grouping.var)>1) {
            grouping <- paste2(grouping.var)
        } else {
            grouping <- unlist(grouping.var)
        } 
    } 
    text <- as.character(text.var)
    DF <- na.omit(data.frame(group = grouping, text.var = text, 
        stringsAsFactors = FALSE))
    if (rm.incomplete) {
        DF <- end_inc(dataframe = DF, text.var = text.var, ...)
    }    
    DF$word.count <- word.count(DF$text.var, missing = 0)
    i <- as.data.frame(table(DF$group))
    DF$group <- DF$group[ , drop=TRUE]
    DF$tot.n.sent <- 1:nrow(DF)
    DF <- DF[with(DF, order(group, DF$tot.n.sent)), ]
    DF$character.count <- character.count(DF$text.var)
    DF2 <- aggregate(word.count ~ group, DF, sum)
    DF2$sentence.count <- as.data.frame(table(DF$group))$Freq
    DF2$character.count <- aggregate(character.count ~ 
        group, DF, sum)$character.count 
    ari <- function(tse, tw, tc) 4.71*(tc/tw) + .5*(tw/tse) - 21.43
    DF2$Automated_Readability_Index <- round(with(DF2, 
        ari(tse = sentence.count, tc = character.count, tw = word.count)), 
        digits = 1)
    names(DF2)[1] <- G
    DF2
}

#' Coleman Liau Readability
#' 
#' \code{coleman_liau} - Apply Coleman Liau Index to transcript(s) by zero or 
#' more grouping variable(s).
#' 
#' @rdname Readability
#' @export
coleman_liau <-
function(text.var, grouping.var = NULL, rm.incomplete = FALSE, ...) {
    if(is.null(grouping.var)) {
        G <- "all"
    } else {
        if (is.list(grouping.var)) {
            m <- unlist(as.character(substitute(grouping.var))[-1])
            m <- sapply(strsplit(m, "$", fixed=TRUE), function(x) {
                    x[length(x)]
                }
            )
            G <- paste(m, collapse="&")
        } else {
            G <- as.character(substitute(grouping.var))
            G <- G[length(G)]
        }
    }
    if(is.null(grouping.var)){
        grouping <- rep("all", length(text.var))
    } else {
        if (is.list(grouping.var) & length(grouping.var)>1) {
            grouping <- paste2(grouping.var)
        } else {
            grouping <- unlist(grouping.var)
        } 
    } 
    text <- as.character(text.var)
    DF <- na.omit(data.frame(group = grouping, text.var = text, 
        stringsAsFactors = FALSE))
    if (rm.incomplete) {
        DF <- end_inc(dataframe = DF, text.var = text.var, ...)
    }
    DF$word.count <- word.count(DF$text.var, missing = 0, digit.remove = FALSE)
    i <- as.data.frame(table(DF$group))
    DF$group <- DF$group[ , drop=TRUE]
    DF$tot.n.sent <- 1:nrow(DF)
    DF <- DF[with(DF, order(group, DF$tot.n.sent)), ]
    DF$character.count <- character.count(DF$text.var, 
        apostrophe.remove = FALSE, digit.remove = FALSE)
    DF2 <- aggregate(word.count ~ group, DF, sum)
    DF2$sentence.count <- as.data.frame(table(DF$group))$Freq
    DF2$character.count <- aggregate(character.count ~ 
        group, DF, sum)$character.count 
    clf <- function(tse, tw, tc) (.0588*((100*tc)/tw)) - 
      (.296*((100*tse)/tw) ) - 15.8
    DF2$Coleman_Liau <- round(with(DF2, clf(tse = sentence.count, 
        tc = character.count, tw = word.count)), digits = 1)
    names(DF2)[1] <- G
    DF2
}


#' SMOG Readability
#' 
#' \code{SMOG} - Apply SMOG Readability to transcript(s) by zero or more grouping variable(s).
#' 
#' @rdname Readability
#' @param output A character vector character string indicating output type. 
#' One of "valid" (default and congruent with McLaughlin's intent) or "all". 
#' @export
SMOG <-
function(text.var, grouping.var = NULL, output = "valid", 
    rm.incomplete = FALSE, ...) {
    group <- NULL
    if(is.null(grouping.var)) {
        G <- "all"
    } else {
        if (is.list(grouping.var)) {
            m <- unlist(as.character(substitute(grouping.var))[-1])
            m <- sapply(strsplit(m, "$", fixed=TRUE), function(x) {
                    x[length(x)]
                }
            )
            G <- paste(m, collapse="&")
        } else {
            G <- as.character(substitute(grouping.var))
            G <- G[length(G)]
        }
    }
    if(is.null(grouping.var)){
        grouping <- rep("all", length(text.var))
    } else {
        if (is.list(grouping.var) & length(grouping.var)>1) {
            grouping <- paste2(grouping.var)
        } else {
            grouping <- unlist(grouping.var)
        } 
    } 
    text <- as.character(text.var)
    DF <- na.omit(data.frame(group = grouping, text.var = text, 
        stringsAsFactors = FALSE))
    if (rm.incomplete) {
        DF <- end_inc(dataframe = DF, text.var = text.var, ...)
    }
    DF$word.count <- word.count(DF$text.var, missing = 0)
    i <- as.data.frame(table(DF$group))
    if (output == "valid") {
        DF <- subset(DF, group%in%as.character(i[i$Freq > 29, ][,'Var1']))
        if (nrow(DF) == 0) {
            stop("Not enough sentences for valid output.")
        }
    }
    DF$group <- DF$group[ , drop=TRUE]
    DF$tot.n.sent <- 1:nrow(DF)
    DF <- DF[with(DF, order(group, DF$tot.n.sent)), ]
    DF$polysyllable.count <- polysyllable.sum(DF$text.var)
    DF2 <- aggregate(word.count ~ group, DF, sum)
    DF2$sentence.count <- as.data.frame(table(DF$group))$Freq
    DF2$polysyllable.count <- aggregate(polysyllable.count ~ 
        group, DF, sum)$polysyllable.count   
    smog <- function(tse, tpsy) 1.043 * sqrt(tpsy * (30/tse)) + 3.1291 
    DF2$SMOG <- round(with(DF2, smog(tse = sentence.count, 
        tpsy = polysyllable.count)), digits = 1)
    DF2$validity <- ifelse(DF2$sentence.count < 30, "n < 30", "valid")
    if(output == "valid") DF2$validity <- NULL
    names(DF2)[1] <- G
    DF2
}


#' Flesch-Kincaid Readability
#' 
#' \code{flesch_kincaid} - Flesch-Kincaid Readability to transcript(s) by zero or more 
#' grouping variable(s).
#' 
#' @rdname Readability
#' @export
flesch_kincaid <-
function(text.var, grouping.var = NULL, rm.incomplete = FALSE, ...) {
      if(is.null(grouping.var)) {
        G <- "all"
    } else {
        if (is.list(grouping.var)) {
            m <- unlist(as.character(substitute(grouping.var))[-1])
            m <- sapply(strsplit(m, "$", fixed=TRUE), function(x) {
                    x[length(x)]
                }
            )
            G <- paste(m, collapse="&")
        } else {
            G <- as.character(substitute(grouping.var))
            G <- G[length(G)]
        }
    }
    if(is.null(grouping.var)){
        grouping <- rep("all", length(text.var))
    } else {
        if (is.list(grouping.var) & length(grouping.var)>1) {
            grouping <- paste2(grouping.var)
        } else {
            grouping <- unlist(grouping.var)
        } 
    } 
    text <- as.character(text.var)
    DF <- na.omit(data.frame(group = grouping, text.var = text, 
        stringsAsFactors = FALSE))
    if (rm.incomplete) {
        DF <- end_inc(dataframe = DF, text.var = text.var, ...)
    }
    DF$word.count <- word.count(DF$text.var, missing = 0)
    DF$syllable.count <- syllable.sum(DF$text.var)
    DF$tot.n.sent <- 1:nrow(DF)
    DF <- DF[with(DF, order(group, DF$tot.n.sent)), ]
    DF2 <- aggregate(word.count ~ group, DF, sum)
    DF2$sentence.count <- as.data.frame(table(DF$group))$Freq
    DF2$syllable.count <- aggregate(syllable.count ~ group, DF, 
        sum)$syllable.count  
    fkgl <- function(tw, tse, tsy) 0.39 * (tw/tse) + 11.8 * (tsy/tw) - 15.59
    fre <- function(tw, tse, tsy) 206.835 - 1.015 * (tw/tse) - 84.6 * (tsy/tw)
    DF2$FK_grd.lvl <- round(with(DF2, fkgl(tw = word.count, 
        tse = sentence.count, tsy = syllable.count)), digits = 1)
    DF2$FK_read.ease <- round(with(DF2, fre(tw = word.count, 
        tse = sentence.count, tsy = syllable.count)), digits = 3)
    names(DF2)[1] <- G
    DF2
}


#' Fry Readability
#' 
#' \code{fry} - Apply Fry Readability to transcript(s) by zero or more 
#' grouping variable(s).
#' 
#' @rdname Readability
#' @param labels  A character vector character string indicating output type. 
#' One of \code{"automatic"} (default; adds labels automatically) or 
#' \code{"click"} (interactive). 
#' @export
fry <-
function(text.var, grouping.var = NULL, labels = "automatic", 
    rm.incomplete = FALSE, ...) {
    read.gr <- group <- NULL  
    if(is.null(grouping.var)) {
        G <- "all"
    } else {
        if (is.list(grouping.var)) {
            m <- unlist(as.character(substitute(grouping.var))[-1])
            m <- sapply(strsplit(m, "$", fixed=TRUE), function(x) {
                    x[length(x)]
                }
            )
            G <- paste(m, collapse="&")
        } else {
            G <- as.character(substitute(grouping.var))
            G <- G[length(G)]
        }
    }
    if(is.null(grouping.var)){
        grouping <- rep("all", length(text.var))
    } else {
        if (is.list(grouping.var) & length(grouping.var)>1) {
            grouping <- paste2(grouping.var)
        } else {
            grouping <- unlist(grouping.var)
        } 
    } 
    text <- as.character(text.var)
    DF <- na.omit(data.frame(group = grouping, text.var = text, 
        stringsAsFactors = FALSE))
    if (rm.incomplete) {
        DF <- end_inc(dataframe = DF, text.var = text.var, ...)
    }
    DF$word.count <- word.count(DF$text.var, missing = 0)
    DF$tot.n.sent <- 1:nrow(DF)
    DF <- DF[with(DF, order(group, DF$tot.n.sent)), ]
    DF$read.gr <- unlist(by(DF$word.count, DF$group, partition))
    LIST <- names(which(tapply(DF$word.count, DF$group, function(x) {
        max(cumsum(x))
    }) >= 300))
    DF2 <- subset(DF, group %in% LIST & read.gr != "NA")
    DF2$group <- DF2$group[, drop = TRUE]
    DF2$sub <- with(DF2, paste(group, read.gr, sep = "_"))
    DF2b <- DF2[order(DF2$sub), ]
    DF2b$cumsum.gr <- unlist(tapply(DF2$word.count, DF2$sub, cumsum))
    DF2 <- DF2b[order(DF2b$group, DF2b$tot.n.sent), ]
    DF2$wo.last <- DF2$cumsum.gr - DF2$word.count
    DF2$hun.word <- 100 - DF2$wo.last
    DF2$frac.sent <- ifelse(DF2$hun.word/DF2$word.count > 1, 
        1, round(DF2$hun.word/DF2$word.count, digits = 3))   
    DF3b <- unique(data.frame(sub = DF2$sub, group = DF2$group))
    DF3 <- data.frame(sub = DF2$sub, group = DF2$group, 
        text.var = DF2$text.var, frac.sent = DF2$frac.sent)
    CHOICES <- tapply(as.character(DF3b$sub), DF3b$group, function(x) {
            sample(x, 3, replace = FALSE)
        }
    )
    CHOICES <- as.character(unlist(CHOICES))
    DF4 <- DF3[which(DF3$sub %in% CHOICES), ]
    DF4$sub <- DF4$sub[, drop = TRUE]
    FUN <- function(x) paste(as.character(unlist(x)), collapse = " ")
    DF5 <- aggregate(text.var ~ sub + group, DF4, FUN)
    sent.per.100 <- as.data.frame(tapply(DF2$frac.sent, DF2$sub, sum))
    names(sent.per.100) <- "x"
    DF5$sent.per.100 <- sent.per.100[as.character(DF5$sub), "x"] 
    hun.grab <- function(x) paste(unblanker(unlist(word.split(
        reducer(unlist(strip(x))))))[1:100], collapse = " ")
    DF5$syll.count <- syllable.sum(lapply(DF5$text.var, hun.grab))
    DF6 <- aggregate(syll.count ~ group, DF5, mean)
    DF6$ave.sent.per.100 <- aggregate(sent.per.100 ~ group, DF5, mean)[, 2]
    suppressWarnings(plot(1, 1, xlim = c(108, 182), ylim = c(2, 
        25), axes = FALSE, type = "n", 
        xlab = "Average number of syllables per 100 words", 
        ylab = "Average number of sentences per 100 words", 
        main = "Fry Graph for Estimating Reading Ages (grade level)", 
        xaxs = "i", yaxs = "i"))
    axis(1, at = 108:182, labels = TRUE)
    axis(1, at = seq(108, 182, by = 4), tcl = -1.1, labels = FALSE)
    axis(2, at = 2:25, labels = TRUE)
    axis(4, at = 2:25, labels = TRUE)
    grid(nx = 74, ny = 46, lty = "solid", col = "gold")
    grid(nx = 37, ny = 23, lty = "solid", col = "gray65")
    box()
    segments(108, 9.97, 133.9, 25, lwd = 2, col = "darkgreen")
    segments(108, 7.7, 142.1, 25, lwd = 2, col = "darkgreen")
    segments(108, 6.4, 147.8, 25, lwd = 2, col = "darkgreen")
    segments(108, 5.61, 155, 25, lwd = 2, col = "darkgreen")
    segments(108, 5.1, 160, 25, lwd = 2, col = "darkgreen")
    segments(108, 3.8, 160.8, 25, lwd = 2, col = "darkgreen")
    segments(111.5, 2, 167.5, 25, lwd = 2, col = "darkgreen")
    segments(123, 2, 169, 25, lwd = 2, col = "darkgreen")
    segments(136.3, 2, 169, 25, lwd = 2, col = "darkgreen")
    segments(143.3, 2, 173.5, 25, lwd = 2, col = "darkgreen")
    segments(148.15, 2, 178.2, 25, lwd = 2, col = "darkgreen")
    segments(155.6, 2, 178.5, 25, lwd = 2, col = "darkgreen")
    segments(161.9, 2, 178.4, 25, lwd = 2, col = "darkgreen")
    segments(168.2, 2, 178.2, 25, lwd = 2, col = "darkgreen")
    segments(173.9, 2, 178.7, 25, lwd = 2, col = "darkgreen")
    segments(179.1, 2, 181.25, 25, lwd = 2, col = "darkgreen")
    x1 <- c(108, 108, 128)
    y1 <- c(4.2, 2, 2)
    polygon(x1, y1, col = "darkblue", border = "darkblue")
    x2 <- c(143, 145, 150, 152, 155, 158, 162, 166, 172, 180, 181, 182, 182)
    y2 <- c(25, 18.3, 14.3, 12.5, 11.1, 10, 9.1, 8.3, 7.7, 7.55, 7.5, 7.5, 25)
    polygon(x2, y2, col = "darkblue", border = "darkblue")  
    text(120, 22.2, 1, col = "darkgreen", cex = 1.25)
    text(124, 17.9, 2, col = "darkgreen", cex = 1.25)
    text(126, 15.9, 3, col = "darkgreen", cex = 1.25)
    text(127, 14.4, 4, col = "darkgreen", cex = 1.25)
    text(128, 13.3, 5, col = "darkgreen", cex = 1.25)
    text(110.7, 5.5, 6, col = "darkgreen", cex = 1.25)
    text(115, 4.88, 7, col = "darkgreen", cex = 1.25)
    text(121, 3.78, 8, col = "darkgreen", cex = 1.25)
    text(132, 3.5, 9, col = "darkgreen", cex = 1.25)
    text(142, 3.5, 10, col = "darkgreen", cex = 1.25)
    text(147.7, 3.5, 11, col = "darkgreen", cex = 1.25)
    text(153.7, 3.5, 12, col = "darkgreen", cex = 1.25)
    text(160, 3.5, 13, col = "darkgreen", cex = 1.25)
    text(166.4, 3.5, 14, col = "darkgreen", cex = 1.25)
    text(171, 3.5, 15, col = "darkgreen", cex = 1.25)
    text(176.7, 3.5, 16, col = "darkgreen", cex = 1.25)
    text(180.7, 3.5, 17, col = "darkgreen", cex = 1.25)
    with(DF6, points(syll.count, ave.sent.per.100, pch = 19, cex = 1, 
        col = "red"))
    with(DF6, points(syll.count, ave.sent.per.100, pch = 3, cex = 3, 
        col = "black"))   
    switch(labels, 
        click = {with(DF6, identify(syll.count, ave.sent.per.100, group))}, 
        automatic = {with(DF6, text(syll.count, ave.sent.per.100, 
        labels = group, adj = c(1, -0.5)))}
    )      
    names(DF6) <- c(G, "ave.syll.per.100", "ave.sent.per.100")
    invisible(list(SENTENCES_USED = DF5, SENTENCE_AVERAGES = DF6))
}


#' Linsear Write Readability
#' 
#' \code{linsear_write} - Apply Linsear Write Readability to transcript(s) by 
#' zero or more grouping variable(s).
#' @rdname Readability
#' @export 
linsear_write <-
function(text.var, grouping.var = NULL, rm.incomplete = FALSE, ...) {
    read.gr <- group <- NULL
    if(is.null(grouping.var)) {
        G <- "all"
    } else {
        if (is.list(grouping.var)) {
            m <- unlist(as.character(substitute(grouping.var))[-1])
            m <- sapply(strsplit(m, "$", fixed=TRUE), function(x) {
                    x[length(x)]
                }
            )
            G <- paste(m, collapse="&")
        } else {
            G <- as.character(substitute(grouping.var))
            G <- G[length(G)]
        }
    }
    if(is.null(grouping.var)){
        grouping <- rep("all", length(text.var))
    } else {
        if (is.list(grouping.var) & length(grouping.var)>1) {
            grouping <- paste2(grouping.var)
        } else {
            grouping <- unlist(grouping.var)
        } 
    } 
    text <- as.character(text.var)
    DF <- na.omit(data.frame(group = grouping, text.var = text, 
        stringsAsFactors = FALSE))
    if (rm.incomplete) {
        DF <- end_inc(dataframe = DF, text.var = text.var, ...)
    }
    DF$word.count <- word.count(DF$text.var)
    DF$tot.n.sent <- 1:nrow(DF)
    DF <- DF[with(DF, order(group, DF$tot.n.sent)), ]
    DF$read.gr <- unlist(by(DF$word.count, DF$group, partition))
    LIST <- names(which(tapply(DF$word.count, DF$group, function(x) {
            max(cumsum(x))
        }
    ) >= 100))
    DF2 <- subset(DF, group %in% LIST & read.gr != "NA")
    DF2$group <- DF2$group[, drop = TRUE]
    DF2$sub <- with(DF2, paste(group, read.gr, sep = "_"))
    DF2b <- DF2[order(DF2$sub), ]
    DF2b$cumsum.gr <- unlist(tapply(DF2$word.count, DF2$sub, cumsum))
    DF2 <- DF2b[order(DF2b$group, DF2b$tot.n.sent), ]
    DF2$wo.last <- DF2$cumsum.gr - DF2$word.count
    DF2$hun.word <- 100 - DF2$wo.last
    DF2$frac.sent <- ifelse(DF2$hun.word/DF2$word.count > 1, 1, 
        round(DF2$hun.word/DF2$word.count, digits = 3))  
    DF3b <- unique(data.frame(sub = DF2$sub, group = DF2$group))
    DF3 <- data.frame(sub = DF2$sub, group = DF2$group, text.var = 
        DF2$text.var, frac.sent = DF2$frac.sent)
    CHOICES <- tapply(as.character(DF3b$sub), DF3b$group, function(x) {
            sample(x, 1, replace = FALSE)
        }
    )
    CHOICES <- as.character(unlist(CHOICES))
    DF4 <- DF3[which(DF3$sub %in% CHOICES), ]
    DF4$sub <- DF4$sub[, drop = TRUE]
    FUN <- function(x) paste(as.character(unlist(x)), collapse = " ")
    DF5 <- aggregate(text.var ~ sub + group, DF4, FUN)
    sent.per.100 <- as.data.frame(tapply(DF2$frac.sent, DF2$sub, sum))
    names(sent.per.100) <- "x"
    DF5$sent.per.100 <- sent.per.100[as.character(DF5$sub), "x"]
    hun.grab <- function(x) paste(unblanker(unlist(word.split(reducer(
        unlist(strip(x))))))[1:100], collapse = " ")
    DF5$SYL.LIST <- lapply(DF5$text.var, function(x) unlist(syllable.count(
        hun.grab(x))$syllables))
    DF5$hard_easy_sum <- unlist(lapply(DF5$SYL.LIST, function(x) {
          sum(ifelse(x >= 3, 3, 1))
    }))
    DF5$HE_tsent_ratio <- unlist(DF5$hard_easy_sum)/DF5$sent.per.100
    DF5$Linsear_Write <- ifelse(DF5$sent.per.100 > 20, 
        round(DF5$HE_tsent_ratio/2, digits = 2), 
        round((DF5$HE_tsent_ratio - 2)/2, digits = 2))
    DF5 <- DF5[, c(2, 4, 6, 8)]
    names(DF5) <- c(G, "sent.per.100", "hard_easy_sum", "Linsear_Write")
    DF5
}